all_solutions
listlengths
1
38
answer
stringlengths
7
23
bucket
stringclasses
1 value
canonical_solution
stringlengths
7
23
config
dict
difficulty
stringclasses
1 value
difficulty_proxy
float64
-3.64
0
generator_seed
int64
21k
21k
info
dict
n_operands
int64
4
4
num_solutions
int64
1
38
operand_range
listlengths
2
2
operands
listlengths
4
4
pass10
dict
problem_id
stringlengths
29
29
question
stringlengths
388
392
split
stringclasses
1 value
subset
stringclasses
1 value
target
int64
10
98
[ "(24 + 23)", "((24 * 1) + 23)", "((24 / 1) + 23)", "(1 * (24 + 23))", "((24 + 23) / 1)", "(24 + (1 * 23))", "(24 + (23 / 1))" ]
(24 + 23)
bucket_a
(24 + 23)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "(24 + 23)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required...
4
7
[ 1, 25 ]
[ 10, 24, 1, 23 ]
{ "completion_tokens": 3839, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 3, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01730
Using the numbers [10, 24, 1, 23], write an arithmetic expression that equals 47. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
47
[ "(17 * 4)", "((17 * 1) * 4)", "((17 / 1) * 4)", "(1 * (17 * 4))", "((17 * 4) / 1)", "(17 * (1 * 4))", "(17 * (4 / 1))", "((17 - 1) + (13 * 4))", "((17 + (13 * 4)) - 1)", "(17 + ((13 * 4) - 1))" ]
(17 * (4 / 1))
bucket_a
(17 * (4 / 1))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.302585
21,001
{ "canonical_solution": "(17 * (4 / 1))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
10
[ 1, 25 ]
[ 13, 17, 1, 4 ]
{ "completion_tokens": 3467, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 1, "no_box": 1, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01737
Using the numbers [13, 17, 1, 4], write an arithmetic expression that equals 68. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
68
[ "(22 - ((3 + 13) / 8))", "(22 - ((13 - 3) - 8))", "(22 - (13 - (3 + 8)))", "(22 - ((13 - 8) - 3))", "((3 + 22) - (13 - 8))", "((3 + 8) + (22 - 13))", "(8 + (22 - (13 - 3)))", "(8 + ((3 + 22) - 13))", "(8 + (3 + (22 - 13)))", "((8 + 22) - (13 - 3))", "(((3 + 8) + 22) - 13)", "((8 + (3 + 22)) - ...
(3 + ((22 + 8) - 13))
bucket_a
(3 + ((22 + 8) - 13))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "(3 + ((22 + 8) - 13))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 3, 13, 8, 22 ]
{ "completion_tokens": 2471, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "parse_error": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01743
Using the numbers [3, 13, 8, 22], write an arithmetic expression that equals 20. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
20
[ "(((25 + 10) + 19) + 24)", "((10 + (25 + 19)) + 24)", "((25 + (10 + 19)) + 24)", "((10 + 19) + (25 + 24))", "((25 + 19) + (10 + 24))", "(19 + ((25 + 10) + 24))", "(19 + (10 + (25 + 24)))", "(19 + (25 + (10 + 24)))", "((25 + 10) + (19 + 24))", "(10 + ((25 + 19) + 24))", "(10 + (19 + (25 + 24)))",...
((10 + 19) + (25 + 24))
bucket_a
((10 + 19) + (25 + 24))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((10 + 19) + (25 + 24))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1,...
4
15
[ 1, 25 ]
[ 25, 10, 19, 24 ]
{ "completion_tokens": 3614, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01758
Using the numbers [25, 10, 19, 24], write an arithmetic expression that equals 78. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
78
[ "(14 - 4)", "(((14 + 9) - 9) - 4)", "(((14 * 9) / 9) - 4)", "(((14 - 9) + 9) - 4)", "((9 + (14 - 9)) - 4)", "((14 * (9 / 9)) - 4)", "((14 / (9 / 9)) - 4)", "((9 / 9) * (14 - 4))", "((14 - 4) / (9 / 9))", "((14 + 9) - (9 + 4))", "((14 - 9) + (9 - 4))", "(((14 + 9) - 4) - 9)", "((9 + (14 - 4))...
((14 - 4) * (9 / 9))
bucket_a
((14 - 4) * (9 / 9))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.135494
21,001
{ "canonical_solution": "((14 - 4) * (9 / 9))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
23
[ 1, 25 ]
[ 14, 9, 9, 4 ]
{ "completion_tokens": 1915, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_integer_intermediate": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01764
Using the numbers [14, 9, 9, 4], write an arithmetic expression that equals 10. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
10
[ "(10 + (19 * 3))", "((19 + 18) + (10 * 3))", "(18 + (19 + (10 * 3)))", "(19 + (18 + (10 * 3)))" ]
((19 * 3) + 10)
bucket_a
((19 * 3) + 10)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.386294
21,001
{ "canonical_solution": "((19 * 3) + 10)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
4
[ 1, 25 ]
[ 19, 18, 10, 3 ]
{ "completion_tokens": 4643, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01771
Using the numbers [19, 18, 10, 3], write an arithmetic expression that equals 67. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
67
[ "(((18 - 10) * 6) + 4)", "((10 + (18 / 6)) * 4)", "((18 - 6) + (10 * 4))", "((18 + (10 * 4)) - 6)", "((18 + 10) + (6 * 4))", "(10 + (18 + (6 * 4)))", "(18 + ((10 * 4) - 6))", "(18 + (10 + (6 * 4)))" ]
(((4 * 10) - 6) + 18)
bucket_a
(((4 * 10) - 6) + 18)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.079442
21,001
{ "canonical_solution": "(((4 * 10) - 6) + 18)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
8
[ 1, 25 ]
[ 18, 10, 6, 4 ]
{ "completion_tokens": 4501, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01782
Using the numbers [18, 10, 6, 4], write an arithmetic expression that equals 52. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
52
[ "((7 + 23) + 15)", "(23 + (7 + 15))", "(7 + (23 + 15))", "(((23 - 7) + 15) + 14)", "((23 + (15 - 7)) + 14)", "(((23 + 15) - 7) + 14)", "((23 + 15) + (14 - 7))", "((15 - 7) + (23 + 14))", "(15 + ((23 - 7) + 14))", "(15 + (23 + (14 - 7)))", "(15 + ((23 + 14) - 7))", "((23 - 7) + (15 + 14))", "...
((14 + (23 - 7)) + 15)
bucket_a
((14 + (23 - 7)) + 15)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((14 + (23 - 7)) + 15)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
18
[ 1, 25 ]
[ 7, 23, 15, 14 ]
{ "completion_tokens": 4864, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01794
Using the numbers [7, 23, 15, 14], write an arithmetic expression that equals 45. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
45
[ "(((13 + 23) + 20) - 3)", "((23 + (13 + 20)) - 3)", "((13 + (23 + 20)) - 3)", "((23 + 20) + (13 - 3))", "((13 + 20) + (23 - 3))", "(20 + ((13 + 23) - 3))", "(20 + (23 + (13 - 3)))", "(20 + (13 + (23 - 3)))", "((13 + 23) + (20 - 3))", "(23 + ((13 + 20) - 3))", "(23 + (20 + (13 - 3)))", "(23 + (...
(13 + ((23 + 20) - 3))
bucket_a
(13 + ((23 + 20) - 3))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(13 + ((23 + 20) - 3))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 13, 23, 20, 3 ]
{ "completion_tokens": 2620, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "non_positive_intermediate": 2, "parse_error": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01796
Using the numbers [13, 23, 20, 3], write an arithmetic expression that equals 53. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
53
[ "(6 * 8)", "((1 * 6) * 8)", "((6 / 1) * 8)", "(6 * (1 * 8))", "(6 * (8 / 1))", "(1 * (6 * 8))", "((6 * 8) / 1)" ]
(6 * 8)
bucket_a
(6 * 8)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "(6 * 8)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_a...
4
7
[ 1, 25 ]
[ 18, 1, 6, 8 ]
{ "completion_tokens": 4160, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "no_box": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01805
Using the numbers [18, 1, 6, 8], write an arithmetic expression that equals 48. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
48
[ "(25 - 8)", "(((25 - 8) + 14) - 14)", "(((25 + 14) - 8) - 14)", "((25 + (14 - 8)) - 14)", "(((25 - 8) * 14) / 14)", "(((25 - 8) - 14) + 14)", "(((25 - 14) - 8) + 14)", "((25 - (8 + 14)) + 14)", "((25 + 14) - (8 + 14))", "((14 - 8) + (25 - 14))", "((25 - 14) + (14 - 8))", "(14 + ((25 - 8) - 14)...
((14 + 25) - (14 + 8))
bucket_a
((14 + 25) - (14 + 8))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((14 + 25) - (14 + 8))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
26
[ 1, 25 ]
[ 25, 8, 14, 14 ]
{ "completion_tokens": 3889, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "non_integer_intermediate": 1, "non_positive_intermediate": 3 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01808
Using the numbers [25, 8, 14, 14], write an arithmetic expression that equals 17. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
17
[ "22", "(5 + 17)", "(((22 + 12) + 5) - 17)", "((12 + (22 + 5)) - 17)", "((22 + (12 + 5)) - 17)", "(((22 - 12) - 5) + 17)", "(((22 - 5) - 12) + 17)", "((22 - (12 + 5)) + 17)", "((22 * (12 + 5)) / 17)", "((22 + 17) - (12 + 5))", "((22 * 17) / (12 + 5))", "((12 + 5) + (22 - 17))", "((22 + 5) - (...
(5 + 17)
bucket_a
(5 + 17)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.295837
21,001
{ "canonical_solution": "(5 + 17)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
27
[ 1, 25 ]
[ 22, 12, 5, 17 ]
{ "completion_tokens": 1957, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01822
Using the numbers [22, 12, 5, 17], write an arithmetic expression that equals 22. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
22
[ "((8 + 11) - 6)", "(11 + (8 - 6))", "(8 + (11 - 6))", "((11 - 8) + 10)", "(11 + (10 - 8))", "((11 + 10) - 8)", "((6 - (11 - 8)) + 10)", "(((8 + 6) - 11) + 10)", "((8 - (11 - 6)) + 10)", "((8 + 10) - (11 - 6))", "((8 + 6) - (11 - 10))", "(((8 * 11) - 10) / 6)", "(6 + (10 - (11 - 8)))", "(6 ...
(((8 + 10) + 6) - 11)
bucket_a
(((8 + 10) + 6) - 11)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.218876
21,001
{ "canonical_solution": "(((8 + 10) + 6) - 11)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
25
[ 1, 25 ]
[ 8, 11, 6, 10 ]
{ "completion_tokens": 3794, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 2, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01825
Using the numbers [8, 11, 6, 10], write an arithmetic expression that equals 13. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
13
[ "(18 - 8)", "(8 + 2)", "10", "((18 + 2) - 10)", "(2 + (18 - 10))", "(18 - (10 - 2))", "(((18 - 8) * 2) - 10)", "(((18 / 2) - 8) * 10)", "(10 / ((18 / 2) - 8))", "((18 + 2) / (10 - 8))", "(((18 - 8) + 10) / 2)", "((18 + (10 - 8)) / 2)", "(((18 + 10) - 8) / 2)", "((2 * 10) - (18 - 8))", "(...
(8 + 2)
bucket_a
(8 + 2)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.833213
21,001
{ "canonical_solution": "(8 + 2)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_a...
4
17
[ 1, 25 ]
[ 8, 18, 2, 10 ]
{ "completion_tokens": 4487, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 1, "parse_error": 4, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01843
Using the numbers [8, 18, 2, 10], write an arithmetic expression that equals 10. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
10
[ "(3 * 25)", "(((11 * 3) + 25) + 17)", "(25 * (17 - (11 + 3)))", "(25 * ((17 - 11) - 3))", "(25 * ((17 - 3) - 11))", "(25 + ((11 * 3) + 17))", "((11 * 3) + (25 + 17))" ]
(3 * 25)
bucket_a
(3 * 25)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "(3 * 25)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
7
[ 1, 25 ]
[ 11, 3, 25, 17 ]
{ "completion_tokens": 3857, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01847
Using the numbers [11, 3, 25, 17], write an arithmetic expression that equals 75. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
75
[ "(7 + 21)", "((1 * 7) + 21)", "((7 / 1) + 21)", "(7 + (1 * 21))", "(7 + (21 / 1))", "(1 * (7 + 21))", "((7 + 21) / 1)", "((15 - (1 + 7)) + 21)", "(((15 - 1) - 7) + 21)", "(((15 - 7) - 1) + 21)", "((15 - 7) + (21 - 1))", "((15 - 1) + (21 - 7))", "(15 + (21 - (1 + 7)))", "(15 + ((21 - 1) - 7...
(1 * (21 + 7))
bucket_a
(1 * (21 + 7))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.091042
21,001
{ "canonical_solution": "(1 * (21 + 7))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
22
[ 1, 25 ]
[ 1, 7, 15, 21 ]
{ "completion_tokens": 2127, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01870
Using the numbers [1, 7, 15, 21], write an arithmetic expression that equals 28. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
28
[ "(7 * 8)", "((7 * 1) * 8)", "((7 / 1) * 8)", "(1 * (7 * 8))", "((7 * 8) / 1)", "(7 * (1 * 8))", "(7 * (8 / 1))", "((8 - 1) * 8)", "(8 * (8 - 1))", "(((7 + 1) * 8) - 8)", "(((7 - 1) * 8) + 8)", "(8 + ((7 - 1) * 8))", "((8 * 8) - (7 + 1))", "(((8 * 8) - 7) - 1)", "(((8 * 8) - 1) - 7)" ]
(8 * 7)
bucket_a
(8 * 7)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(8 * 7)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_a...
4
15
[ 1, 25 ]
[ 7, 1, 8, 8 ]
{ "completion_tokens": 3544, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "no_box": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01876
Using the numbers [7, 1, 8, 8], write an arithmetic expression that equals 56. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ex...
train
easy
56
[ "((16 + 18) - 1)", "(18 + (16 - 1))", "(16 + (18 - 1))", "((16 + 1) + 16)", "(1 + (16 + 16))", "(16 + (1 + 16))", "((18 - 1) + 16)", "((18 + 16) - 1)", "(1 + ((18 - 16) * 16))", "(1 + (16 * (18 - 16)))" ]
((16 + 1) + 16)
bucket_a
((16 + 1) + 16)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.302585
21,001
{ "canonical_solution": "((16 + 1) + 16)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
10
[ 1, 25 ]
[ 16, 18, 1, 16 ]
{ "completion_tokens": 2151, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01902
Using the numbers [16, 18, 1, 16], write an arithmetic expression that equals 33. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
33
[ "(4 + (11 * 7))", "((16 / 4) + (11 * 7))", "((11 + 16) * (7 - 4))", "((4 * (16 + 7)) - 11)" ]
((11 * 7) + 4)
bucket_a
((11 * 7) + 4)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.386294
21,001
{ "canonical_solution": "((11 * 7) + 4)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
4
[ 1, 25 ]
[ 11, 4, 16, 7 ]
{ "completion_tokens": 5921, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01904
Using the numbers [11, 4, 16, 7], write an arithmetic expression that equals 81. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
81
[ "(6 * (21 - 16))", "((19 - (16 - 6)) + 21)", "(((6 + 19) - 16) + 21)", "((6 + (19 - 16)) + 21)", "((19 - 16) + (6 + 21))", "((6 + 19) + (21 - 16))", "(19 + (21 - (16 - 6)))", "(19 + ((6 + 21) - 16))", "(19 + (6 + (21 - 16)))", "((19 + 21) - (16 - 6))", "(((6 + 19) + 21) - 16)", "((19 + (6 + 21...
(19 + ((21 - 16) + 6))
bucket_a
(19 + ((21 - 16) + 6))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "(19 + ((21 - 16) + 6))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 6, 16, 19, 21 ]
{ "completion_tokens": 3031, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "parse_error": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01911
Using the numbers [6, 16, 19, 21], write an arithmetic expression that equals 30. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
30
[ "(((22 + 3) + 18) + 4)", "((3 + (22 + 18)) + 4)", "((22 + (3 + 18)) + 4)", "((3 + 18) + (22 + 4))", "((22 + 18) + (3 + 4))", "(18 + ((22 + 3) + 4))", "(18 + (3 + (22 + 4)))", "(18 + (22 + (3 + 4)))", "((22 + 3) + (18 + 4))", "((18 * 4) - (22 + 3))", "(3 + ((22 + 18) + 4))", "(3 + (18 + (22 + 4...
((18 + 3) + (22 + 4))
bucket_a
((18 + 3) + (22 + 4))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((18 + 3) + (22 + 4))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
18
[ 1, 25 ]
[ 22, 3, 18, 4 ]
{ "completion_tokens": 3491, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 3, "no_box": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01916
Using the numbers [22, 3, 18, 4], write an arithmetic expression that equals 47. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
47
[ "(14 - 2)", "(2 * (14 - 8))", "(3 * (8 / 2))", "((3 * 8) / 2)", "(((14 - 2) / 3) + 8)", "(((14 / 2) - 3) + 8)", "((14 + (2 * 3)) - 8)", "((2 * 3) + (14 - 8))", "(((14 * 2) + 8) / 3)", "(3 * ((14 - 2) - 8))", "(3 * ((14 - 8) - 2))", "(3 * (14 - (2 + 8)))", "(((14 / 2) + 8) - 3)", "((3 * 8) ...
((14 - 8) * 2)
bucket_a
((14 - 8) * 2)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.944439
21,001
{ "canonical_solution": "((14 - 8) * 2)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
19
[ 1, 25 ]
[ 14, 2, 3, 8 ]
{ "completion_tokens": 3372, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01929
Using the numbers [14, 2, 3, 8], write an arithmetic expression that equals 12. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
12
[ "((15 + 25) - 3)", "(25 + (15 - 3))", "(15 + (25 - 3))", "(((15 + 25) - 6) + 3)", "((25 + (15 - 6)) + 3)", "((15 + (25 - 6)) + 3)", "((25 - 6) + (15 + 3))", "((15 - 6) + (25 + 3))", "(((15 + 25) + 3) - 6)", "((25 + (15 + 3)) - 6)", "((15 + (25 + 3)) - 6)", "((15 + 25) - (6 - 3))", "(25 + ((1...
((3 + 25) + (15 - 6))
bucket_a
((3 + 25) + (15 - 6))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((3 + 25) + (15 - 6))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
18
[ 1, 25 ]
[ 15, 25, 6, 3 ]
{ "completion_tokens": 4421, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01933
Using the numbers [15, 25, 6, 3], write an arithmetic expression that equals 37. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
37
[ "(15 + 19)", "(((19 - 15) * 9) - 2)" ]
(15 + 19)
bucket_a
(15 + 19)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-0.693147
21,001
{ "canonical_solution": "(15 + 19)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required...
4
2
[ 1, 25 ]
[ 2, 15, 19, 9 ]
{ "completion_tokens": 3662, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "wrong_value": 6 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01938
Using the numbers [2, 15, 19, 9], write an arithmetic expression that equals 34. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
34
[ "(3 + (3 * 18))", "(((3 * 14) - 3) + 18)", "((3 * 14) + (18 - 3))", "(((3 * 14) + 18) - 3)" ]
((18 * 3) + 3)
bucket_a
((18 * 3) + 3)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.386294
21,001
{ "canonical_solution": "((18 * 3) + 3)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
4
[ 1, 25 ]
[ 3, 3, 14, 18 ]
{ "completion_tokens": 4867, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_01947
Using the numbers [3, 3, 14, 18], write an arithmetic expression that equals 57. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
57
[ "((2 + 15) + 2)", "(15 + (2 + 2))", "(15 + (2 * 2))", "(2 + (15 + 2))", "(((1 * 2) + 15) + 2)", "(((2 / 1) + 15) + 2)", "((2 + (1 * 15)) + 2)", "((2 + (15 / 1)) + 2)", "((1 * (2 + 15)) + 2)", "(((2 + 15) / 1) + 2)", "((2 + 15) + (1 * 2))", "((2 + 15) + (2 / 1))", "((1 * 15) + (2 + 2))", "(...
(15 + (2 * (2 / 1)))
bucket_a
(15 + (2 * (2 / 1)))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.367296
21,001
{ "canonical_solution": "(15 + (2 * (2 / 1)))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
29
[ 1, 25 ]
[ 1, 2, 15, 2 ]
{ "completion_tokens": 4750, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01955
Using the numbers [1, 2, 15, 2], write an arithmetic expression that equals 19. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
19
[ "(((25 - 9) + 20) + 15)", "(((25 + 20) - 9) + 15)", "((25 + (20 - 9)) + 15)", "((20 - 9) + (25 + 15))", "((25 + 20) + (15 - 9))", "(20 + ((25 - 9) + 15))", "(20 + ((25 + 15) - 9))", "(20 + (25 + (15 - 9)))", "((25 - 9) + (20 + 15))", "(((25 + 20) + 15) - 9)", "((20 + (25 + 15)) - 9)", "((25 + ...
((20 + 25) + (15 - 9))
bucket_a
((20 + 25) + (15 - 9))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((20 + 25) + (15 - 9))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 25, 9, 20, 15 ]
{ "completion_tokens": 3490, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 2, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01965
Using the numbers [25, 9, 20, 15], write an arithmetic expression that equals 51. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
51
[ "(((24 + 6) - 21) + 25)", "((6 + (24 - 21)) + 25)", "((24 - (21 - 6)) + 25)", "((24 + 25) - (21 - 6))", "((24 - 21) + (6 + 25))", "(((24 + 6) + 25) - 21)", "((6 + (24 + 25)) - 21)", "((24 + (6 + 25)) - 21)", "((24 + 6) + (25 - 21))", "(6 + ((24 - 21) + 25))", "(6 + ((24 + 25) - 21))", "(6 + (2...
((6 + (24 - 21)) + 25)
bucket_a
((6 + (24 - 21)) + 25)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((6 + (24 - 21)) + 25)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 24, 6, 21, 25 ]
{ "completion_tokens": 2779, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "wrong_value": 5 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01976
Using the numbers [24, 6, 21, 25], write an arithmetic expression that equals 34. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
34
[ "(6 + 13)", "(13 + 6)", "(((6 + 13) + 6) - 6)", "((13 + (6 + 6)) - 6)", "((6 + (13 + 6)) - 6)", "(((6 + 13) * 6) / 6)", "((6 * (13 + 6)) / 6)", "(((6 + 13) - 6) + 6)", "(((6 * 13) / 6) + 6)", "(((13 - 6) + 6) + 6)", "((13 * (6 / 6)) + 6)", "((13 / (6 / 6)) + 6)", "(((13 + 6) - 6) + 6)", "(...
(13 + 6)
bucket_a
(13 + 6)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.295837
21,001
{ "canonical_solution": "(13 + 6)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
27
[ 1, 25 ]
[ 6, 13, 6, 6 ]
{ "completion_tokens": 1410, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_01983
Using the numbers [6, 13, 6, 6], write an arithmetic expression that equals 19. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
19
[ "(((13 * 19) - 5) / 11)", "(((19 - 13) + 5) + 11)", "((19 - (13 - 5)) + 11)", "(((19 + 5) - 13) + 11)", "((19 + 5) - (13 - 11))", "((19 + 11) - (13 - 5))", "(5 + ((19 - 13) + 11))", "(5 + (19 - (13 - 11)))", "(5 + ((19 + 11) - 13))", "((19 - 13) + (5 + 11))", "(19 + (11 - (13 - 5)))", "(19 + (...
((19 - (13 - 11)) + 5)
bucket_a
((19 - (13 - 11)) + 5)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "((19 - (13 - 11)) + 5)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 13, 19, 5, 11 ]
{ "completion_tokens": 3589, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01985
Using the numbers [13, 19, 5, 11], write an arithmetic expression that equals 22. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
22
[ "(16 + 1)", "(((22 - 16) + 1) + 10)", "(((22 + 1) - 16) + 10)", "((22 - (16 - 1)) + 10)", "((22 + 10) - (16 - 1))", "((22 + 1) - (16 - 10))", "(1 + ((22 - 16) + 10))", "(1 + ((22 + 10) - 16))", "(1 + (22 - (16 - 10)))", "((22 - 16) + (1 + 10))", "(((22 + 1) + 10) - 16)", "((1 + (22 + 10)) - 16...
(16 + 1)
bucket_a
(16 + 1)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "(16 + 1)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
16
[ 1, 25 ]
[ 22, 16, 1, 10 ]
{ "completion_tokens": 2566, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "non_integer_intermediate": 1, "non_positive_intermediate": 1, "parse_error": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_01998
Using the numbers [22, 16, 1, 10], write an arithmetic expression that equals 17. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
17
[ "(2 * (18 + 1))", "(2 * 19)", "((18 + 1) + 19)", "(1 + (18 + 19))", "(18 + (1 + 19))", "((2 * 1) * 19)", "((2 / 1) * 19)", "(1 * (2 * 19))", "((2 * 19) / 1)", "(2 * (1 * 19))", "(2 * (19 / 1))", "(((18 + 2) - 1) + 19)", "((2 + (18 - 1)) + 19)", "((18 + (2 - 1)) + 19)", "((2 - 1) + (18 + ...
(2 * 19)
bucket_a
(2 * 19)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "(2 * 19)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
26
[ 1, 25 ]
[ 18, 2, 1, 19 ]
{ "completion_tokens": 2832, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02010
Using the numbers [18, 2, 1, 19], write an arithmetic expression that equals 38. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
38
[ "(23 + 12)", "(((23 + 15) + 12) - 15)", "((15 + (23 + 12)) - 15)", "((23 + (15 + 12)) - 15)", "(((23 - 15) + 12) + 15)", "(((23 + 12) - 15) + 15)", "((23 - (15 - 12)) + 15)", "((15 * (23 + 12)) / 15)", "((15 + 12) + (23 - 15))", "((23 + 15) - (15 - 12))", "((23 + 12) * (15 / 15))", "((23 + 12)...
(23 + 12)
bucket_a
(23 + 12)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "(23 + 12)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required...
4
26
[ 1, 25 ]
[ 23, 15, 12, 15 ]
{ "completion_tokens": 2451, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_integer_intermediate": 1, "non_positive_intermediate": 4, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02015
Using the numbers [23, 15, 12, 15], write an arithmetic expression that equals 35. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
35
[ "25", "(11 + 14)", "(((25 - 11) / 14) + 24)", "((14 / (25 - 11)) + 24)", "((11 / (25 - 14)) + 24)", "(((25 - 14) / 11) + 24)", "((25 / (11 + 14)) + 24)", "(((11 + 14) / 25) + 24)", "((11 + 14) * (25 - 24))", "((11 + 14) / (25 - 24))", "(14 + (11 * (25 - 24)))", "(14 + (11 / (25 - 24)))", "(1...
((14 + 11) * (25 - 24))
bucket_a
((14 + 11) * (25 - 24))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.995732
21,001
{ "canonical_solution": "((14 + 11) * (25 - 24))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1,...
4
20
[ 1, 25 ]
[ 25, 11, 14, 24 ]
{ "completion_tokens": 4174, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "non_positive_intermediate": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02017
Using the numbers [25, 11, 14, 24], write an arithmetic expression that equals 25. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
25
[ "(19 + 23)", "(((19 + 4) + 23) - 4)", "((4 + (19 + 23)) - 4)", "((19 + (4 + 23)) - 4)", "(((19 - 4) + 23) + 4)", "(((19 + 23) - 4) + 4)", "((19 + (23 - 4)) + 4)", "((4 * (19 + 23)) / 4)", "((4 + 23) + (19 - 4))", "((23 - 4) + (19 + 4))", "((19 + 23) * (4 / 4))", "((19 + 23) / (4 / 4))", "(23...
((4 + (23 - 4)) + 19)
bucket_a
((4 + (23 - 4)) + 19)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((4 + (23 - 4)) + 19)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
26
[ 1, 25 ]
[ 19, 4, 23, 4 ]
{ "completion_tokens": 2823, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02040
Using the numbers [19, 4, 23, 4], write an arithmetic expression that equals 42. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
42
[ "(12 + 17)", "(((20 / 20) * 12) + 17)", "((12 / (20 / 20)) + 17)", "(((20 + 12) - 20) + 17)", "(((20 * 12) / 20) + 17)", "((20 - (20 - 12)) + 17)", "((20 + 12) - (20 - 17))", "((20 + 17) - (20 - 12))", "(12 + ((20 / 20) * 17))", "(12 + (17 / (20 / 20)))", "(12 + ((20 + 17) - 20))", "(12 + ((20...
((17 + 20) - (20 - 12))
bucket_a
((17 + 20) - (20 - 12))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.091042
21,001
{ "canonical_solution": "((17 + 20) - (20 - 12))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1,...
4
22
[ 1, 25 ]
[ 20, 20, 12, 17 ]
{ "completion_tokens": 2739, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02041
Using the numbers [20, 20, 12, 17], write an arithmetic expression that equals 29. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
29
[ "24", "(((24 + 10) + 3) - 13)", "((10 + (24 + 3)) - 13)", "((24 + (10 + 3)) - 13)", "(((24 - 10) - 3) + 13)", "(((24 - 3) - 10) + 13)", "((24 - (10 + 3)) + 13)", "((24 * (10 + 3)) / 13)", "((24 + 13) - (10 + 3))", "((24 * 13) / (10 + 3))", "((10 + 3) + (24 - 13))", "((24 + 3) - (13 - 10))", ...
((24 * 13) / (3 + 10))
bucket_a
((24 * 13) / (3 + 10))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((24 * 13) / (3 + 10))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
26
[ 1, 25 ]
[ 24, 10, 3, 13 ]
{ "completion_tokens": 2969, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02043
Using the numbers [24, 10, 3, 13], write an arithmetic expression that equals 24. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
24
[ "(10 * (1 + 1))", "((1 + 10) + 9)", "(10 + (1 + 9))", "(1 + (10 + 9))", "((10 + 1) + 9)", "(((1 + 10) * 1) + 9)", "(((1 + 10) / 1) + 9)", "(((1 * 10) + 1) + 9)", "(((10 / 1) + 1) + 9)", "((10 + (1 * 1)) + 9)", "((10 + (1 / 1)) + 9)", "((1 * (10 + 1)) + 9)", "(((10 + 1) / 1) + 9)", "((1 + (...
((9 + (1 * 10)) + 1)
bucket_a
((9 + (1 * 10)) + 1)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.401197
21,001
{ "canonical_solution": "((9 + (1 * 10)) + 1)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
30
[ 1, 25 ]
[ 1, 10, 1, 9 ]
{ "completion_tokens": 2607, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02044
Using the numbers [1, 10, 1, 9], write an arithmetic expression that equals 20. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
20
[ "((12 - 1) * 2)", "((12 - 1) + 11)", "((12 + 11) - 1)", "(12 + (11 - 1))", "(2 * 11)", "((1 * 2) * 11)", "((2 / 1) * 11)", "(2 * (1 * 11))", "(2 * (11 / 1))", "(1 * (2 * 11))", "((2 * 11) / 1)", "(((12 + 1) - 2) + 11)", "((1 + (12 - 2)) + 11)", "((12 - (2 - 1)) + 11)", "((12 + 11) - (2 -...
((1 * 11) * 2)
bucket_a
((1 * 11) * 2)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((1 * 11) * 2)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
26
[ 1, 25 ]
[ 12, 1, 2, 11 ]
{ "completion_tokens": 4164, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02048
Using the numbers [12, 1, 2, 11], write an arithmetic expression that equals 22. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
22
[ "((8 - 5) * 23)", "(((8 * 5) + 23) + 6)", "((5 * (23 - 8)) - 6)", "(23 + ((8 * 5) + 6))", "(23 * (6 - (8 - 5)))", "(23 * (5 - (8 - 6)))", "(23 * ((5 + 6) - 8))", "((8 * 5) + (23 + 6))" ]
(6 + (23 + (8 * 5)))
bucket_a
(6 + (23 + (8 * 5)))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.079442
21,001
{ "canonical_solution": "(6 + (23 + (8 * 5)))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
8
[ 1, 25 ]
[ 8, 5, 23, 6 ]
{ "completion_tokens": 4343, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "no_box": 1, "parse_error": 3, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02052
Using the numbers [8, 5, 23, 6], write an arithmetic expression that equals 69. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
69
[ "(15 + 23)", "(((2 + 15) + 23) - 2)", "((15 + (2 + 23)) - 2)", "((2 + (15 + 23)) - 2)", "(((15 - 2) + 23) + 2)", "((15 + (23 - 2)) + 2)", "(((15 + 23) - 2) + 2)", "((2 * (15 + 23)) / 2)", "(((15 + 23) / 2) * 2)", "((15 + 23) * (2 / 2))", "((15 + 23) / (2 / 2))", "((2 + 23) + (15 - 2))", "((2...
((2 + 23) + (15 - 2))
bucket_a
((2 + 23) + (15 - 2))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((2 + 23) + (15 - 2))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
26
[ 1, 25 ]
[ 2, 15, 23, 2 ]
{ "completion_tokens": 1586, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "parse_error": 3, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02068
Using the numbers [2, 15, 23, 2], write an arithmetic expression that equals 38. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
38
[ "(10 + 5)", "((12 - 10) + 13)", "((12 + 13) - 10)", "(12 + (13 - 10))", "((10 / 5) + 13)", "(5 * (13 - 10))", "((10 + 5) * (13 - 12))", "((10 + 5) / (13 - 12))", "(5 + (10 * (13 - 12)))", "(5 + (10 / (13 - 12)))", "((12 * 10) / (13 - 5))", "(10 + ((12 + 13) / 5))", "(10 + (5 * (13 - 12)))", ...
((12 + 13) - 10)
bucket_a
((12 + 13) - 10)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((12 + 13) - 10)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "r...
4
15
[ 1, 25 ]
[ 12, 10, 5, 13 ]
{ "completion_tokens": 2567, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02076
Using the numbers [12, 10, 5, 13], write an arithmetic expression that equals 15. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
15
[ "(((9 + 24) + 22) - 14)", "((24 + (9 + 22)) - 14)", "((9 + (24 + 22)) - 14)", "((24 + 22) - (14 - 9))", "((9 + 22) + (24 - 14))", "(22 + ((9 + 24) - 14))", "(22 + (24 - (14 - 9)))", "(22 + (9 + (24 - 14)))", "((9 + 24) + (22 - 14))", "(24 + ((9 + 22) - 14))", "(24 + (22 - (14 - 9)))", "(24 + (...
((22 + (24 + 9)) - 14)
bucket_a
((22 + (24 + 9)) - 14)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((22 + (24 + 9)) - 14)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 9, 24, 22, 14 ]
{ "completion_tokens": 2466, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02083
Using the numbers [9, 24, 22, 14], write an arithmetic expression that equals 41. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
41
[ "(19 + 2)", "(23 - 2)", "((19 + 23) / 2)", "(((19 + 23) + 2) - 23)", "((23 + (19 + 2)) - 23)", "((19 + (23 + 2)) - 23)", "(23 - ((23 - 19) - 2))", "(23 - ((23 - 19) / 2))", "(23 - (23 - (19 + 2)))", "(23 - ((23 - 2) - 19))", "((23 * (19 + 2)) / 23)", "((23 + 2) - (23 - 19))", "((19 + 23) - (...
(23 - ((23 - 2) - 19))
bucket_a
(23 - ((23 - 2) - 19))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.332205
21,001
{ "canonical_solution": "(23 - ((23 - 2) - 19))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
28
[ 1, 25 ]
[ 19, 23, 2, 23 ]
{ "completion_tokens": 2139, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 4, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02106
Using the numbers [19, 23, 2, 23], write an arithmetic expression that equals 21. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
21
[ "10", "(18 - (10 - 2))", "((2 + 18) - 10)", "(2 + (18 - 10))" ]
(18 - (10 - 2))
bucket_a
(18 - (10 - 2))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.386294
21,001
{ "canonical_solution": "(18 - (10 - 2))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
4
[ 1, 25 ]
[ 2, 25, 10, 18 ]
{ "completion_tokens": 5026, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "no_box": 1, "parse_error": 4, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02110
Using the numbers [2, 25, 10, 18], write an arithmetic expression that equals 10. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
10
[ "(((9 + 7) + 19) + 23)", "((7 + (9 + 19)) + 23)", "((9 + (7 + 19)) + 23)", "((7 + 19) + (9 + 23))", "((9 + 19) + (7 + 23))", "(19 + ((9 + 7) + 23))", "(19 + (7 + (9 + 23)))", "(19 + (9 + (7 + 23)))", "((9 + 7) + (19 + 23))", "(7 + ((9 + 19) + 23))", "(7 + (19 + (9 + 23)))", "(7 + (9 + (19 + 23...
((7 + 23) + (19 + 9))
bucket_a
((7 + 23) + (19 + 9))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((7 + 23) + (19 + 9))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 9, 7, 19, 23 ]
{ "completion_tokens": 3852, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02128
Using the numbers [9, 7, 19, 23], write an arithmetic expression that equals 58. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
58
[ "(((10 + 13) + 1) + 13)", "((13 + (10 + 1)) + 13)", "((10 + (13 + 1)) + 13)", "((13 + 1) + (10 + 13))", "((10 + 1) + (13 + 13))", "(1 + ((10 + 13) + 13))", "(1 + (13 + (10 + 13)))", "(1 + (10 + (13 + 13)))", "((10 + 13) + (1 + 13))", "(13 + ((10 + 1) + 13))", "(13 + (1 + (10 + 13)))", "(13 + (...
((10 + 1) + (13 + 13))
bucket_a
((10 + 1) + (13 + 13))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((10 + 1) + (13 + 13))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 10, 13, 1, 13 ]
{ "completion_tokens": 4891, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02129
Using the numbers [10, 13, 1, 13], write an arithmetic expression that equals 37. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
37
[ "((8 + 7) + 3)", "(7 + (8 + 3))", "(8 + (7 + 3))", "((16 - (8 - 7)) + 3)", "((7 + (16 - 8)) + 3)", "(((7 + 16) - 8) + 3)", "((7 + 16) - (8 - 3))", "((16 - 8) + (7 + 3))", "(16 + (3 - (8 - 7)))", "(16 + (7 - (8 - 3)))", "(16 + ((7 + 3) - 8))", "(16 + (8 / (7 - 3)))", "((16 + 3) - (8 - 7))", ...
((7 + 3) + 8)
bucket_a
((7 + 3) + 8)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.944439
21,001
{ "canonical_solution": "((7 + 3) + 8)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
19
[ 1, 25 ]
[ 8, 7, 16, 3 ]
{ "completion_tokens": 4004, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_integer_intermediate": 1, "parse_error": 3, "wrong_value": 1 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02142
Using the numbers [8, 7, 16, 3], write an arithmetic expression that equals 18. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
18
[ "(23 * 2)", "((23 + 21) + 2)", "(21 + (23 + 2))", "(23 + (21 + 2))" ]
(2 * 23)
bucket_a
(2 * 23)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.386294
21,001
{ "canonical_solution": "(2 * 23)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
4
[ 1, 25 ]
[ 6, 23, 21, 2 ]
{ "completion_tokens": 3674, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02152
Using the numbers [6, 23, 21, 2], write an arithmetic expression that equals 46. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
46
[ "(25 - (17 - 8))", "((8 + 25) - 17)", "(8 + (25 - 17))", "(17 - 1)", "((25 - 8) - 1)", "(25 - (8 + 1))", "((25 - 1) - 8)", "((25 - (17 - 8)) * 1)", "((25 - (17 - 8)) / 1)", "(((8 + 25) - 17) * 1)", "(((8 + 25) - 17) / 1)", "((8 + (25 - 17)) * 1)", "((8 + (25 - 17)) / 1)", "((25 - 17) + (8 ...
(25 - (1 + 8))
bucket_a
(25 - (1 + 8))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.465736
21,001
{ "canonical_solution": "(25 - (1 + 8))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
32
[ 1, 25 ]
[ 8, 17, 25, 1 ]
{ "completion_tokens": 2443, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 3, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02154
Using the numbers [8, 17, 25, 1], write an arithmetic expression that equals 16. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
16
[ "(9 + 4)", "((8 + 9) - 4)", "(9 + (8 - 4))", "(8 + (9 - 4))", "((9 + 8) - 4)", "(((8 + 9) - 8) + 4)", "(((8 * 9) / 8) + 4)", "(((9 - 8) + 8) + 4)", "((9 * (8 / 8)) + 4)", "((9 / (8 / 8)) + 4)", "(((9 + 8) - 8) + 4)", "(((9 * 8) / 8) + 4)", "((8 + (9 - 8)) + 4)", "((9 + 8) - (8 - 4))", "(...
((9 + 4) * (8 / 8))
bucket_a
((9 + 4) * (8 / 8))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.401197
21,001
{ "canonical_solution": "((9 + 4) * (8 / 8))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
30
[ 1, 25 ]
[ 8, 9, 8, 4 ]
{ "completion_tokens": 2076, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02156
Using the numbers [8, 9, 8, 4], write an arithmetic expression that equals 13. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ex...
train
easy
13
[ "(8 + 8)", "((8 / 4) * 8)", "((8 * 8) / 4)", "(8 * (8 / 4))", "(4 * 4)", "((8 + 4) + 4)", "((8 - 4) * 4)", "(4 + (8 + 4))", "(4 * (8 - 4))", "(8 + (4 + 4))", "((4 + 8) + 4)", "(((8 + 4) + 8) - 4)", "((4 + (8 + 8)) - 4)", "((8 + (4 + 8)) - 4)", "(((8 + 4) - 8) * 4)", "(((8 * 4) / 8) * 4...
(8 * (8 / 4))
bucket_a
(8 * (8 / 4))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.583519
21,001
{ "canonical_solution": "(8 * (8 / 4))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
36
[ 1, 25 ]
[ 8, 4, 8, 4 ]
{ "completion_tokens": 2696, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02159
Using the numbers [8, 4, 8, 4], write an arithmetic expression that equals 16. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ex...
train
easy
16
[ "((6 * 3) * 3)", "(3 * (6 * 3))", "(6 * (3 * 3))", "(3 * (3 + 15))", "((3 + 3) * (15 - 6))", "((6 + 3) + (3 * 15))", "((6 - 3) * (3 + 15))", "(3 * ((6 - 3) + 15))", "(3 * ((6 + 15) - 3))", "(3 * (6 + (15 - 3)))", "(3 + (6 + (3 * 15)))", "(6 * (15 - (3 + 3)))", "(6 * ((15 - 3) - 3))", "(6 +...
((3 * 6) * 3)
bucket_a
((3 * 6) * 3)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.639057
21,001
{ "canonical_solution": "((3 * 6) * 3)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
14
[ 1, 25 ]
[ 6, 3, 3, 15 ]
{ "completion_tokens": 2704, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02167
Using the numbers [6, 3, 3, 15], write an arithmetic expression that equals 54. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
54
[ "(((11 + 4) - 6) + 17)", "((4 + (11 - 6)) + 17)", "((11 - (6 - 4)) + 17)", "((11 + 17) - (6 - 4))", "((11 - 6) + (4 + 17))", "(((11 + 4) + 17) - 6)", "((4 + (11 + 17)) - 6)", "((11 + (4 + 17)) - 6)", "((11 + 4) + (17 - 6))", "(4 + ((11 - 6) + 17))", "(4 + ((11 + 17) - 6))", "(4 + (11 + (17 - 6...
(17 + ((11 - 6) + 4))
bucket_a
(17 + ((11 - 6) + 4))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(17 + ((11 - 6) + 4))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 11, 4, 6, 17 ]
{ "completion_tokens": 4882, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "parse_error": 3, "wrong_value": 1 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02168
Using the numbers [11, 4, 6, 17], write an arithmetic expression that equals 26. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
26
[ "(((11 - 4) + 12) + 24)", "((11 + (12 - 4)) + 24)", "(((11 + 12) - 4) + 24)", "((11 + 12) + (24 - 4))", "((12 - 4) + (11 + 24))", "(12 + ((11 - 4) + 24))", "(12 + (11 + (24 - 4)))", "(12 + ((11 + 24) - 4))", "((11 - 4) + (12 + 24))", "(11 + ((12 - 4) + 24))", "(11 + (12 + (24 - 4)))", "(11 + (...
((11 + (24 - 4)) + 12)
bucket_a
((11 + (24 - 4)) + 12)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((11 + (24 - 4)) + 12)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 4, 11, 12, 24 ]
{ "completion_tokens": 3239, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02178
Using the numbers [4, 11, 12, 24], write an arithmetic expression that equals 43. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
43
[ "(((24 + 20) + 10) + 19)", "((20 + (24 + 10)) + 19)", "((24 + (20 + 10)) + 19)", "((20 + 10) + (24 + 19))", "((24 + 10) + (20 + 19))", "(10 + ((24 + 20) + 19))", "(10 + (20 + (24 + 19)))", "(10 + (24 + (20 + 19)))", "((24 + 20) + (10 + 19))", "(20 + ((24 + 10) + 19))", "(20 + (10 + (24 + 19)))",...
(((19 + 10) + 20) + 24)
bucket_a
(((19 + 10) + 20) + 24)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(((19 + 10) + 20) + 24)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1,...
4
15
[ 1, 25 ]
[ 24, 20, 10, 19 ]
{ "completion_tokens": 2942, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02185
Using the numbers [24, 20, 10, 19], write an arithmetic expression that equals 73. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
73
[ "(((25 - 19) + 23) + 4)", "(((25 + 23) - 19) + 4)", "((25 + (23 - 19)) + 4)", "((23 - 19) + (25 + 4))", "((25 + 23) - (19 - 4))", "(23 + ((25 - 19) + 4))", "(23 + ((25 + 4) - 19))", "(23 + (25 - (19 - 4)))", "((25 - 19) + (23 + 4))", "(((25 + 23) + 4) - 19)", "((23 + (25 + 4)) - 19)", "((25 + ...
(((25 + 4) - 19) + 23)
bucket_a
(((25 + 4) - 19) + 23)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(((25 + 4) - 19) + 23)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 25, 19, 23, 4 ]
{ "completion_tokens": 3688, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02208
Using the numbers [25, 19, 23, 4], write an arithmetic expression that equals 33. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
33
[ "((24 - 11) + 19)", "((24 + 19) - 11)", "(24 + (19 - 11))", "(((24 - 22) + 11) + 19)", "((24 - (22 - 11)) + 19)", "(((24 + 11) - 22) + 19)", "((24 + 11) - (22 - 19))", "((24 + 19) - (22 - 11))", "(11 + ((24 - 22) + 19))", "(11 + (24 - (22 - 19)))", "(11 + ((24 + 19) - 22))", "((24 - 22) + (11 ...
(19 + (24 - (22 - 11)))
bucket_a
(19 + (24 - (22 - 11)))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "(19 + (24 - (22 - 11)))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1,...
4
18
[ 1, 25 ]
[ 22, 24, 11, 19 ]
{ "completion_tokens": 3936, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02214
Using the numbers [22, 24, 11, 19], write an arithmetic expression that equals 32. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
32
[ "(3 + 20)", "(((3 + 8) - 8) + 20)", "(((3 * 8) / 8) + 20)", "((8 - (8 - 3)) + 20)", "((3 * (8 / 8)) + 20)", "((3 / (8 / 8)) + 20)", "((8 / 8) * (3 + 20))", "((3 + 20) / (8 / 8))", "((3 + 8) + (20 - 8))", "((8 + 20) - (8 - 3))", "(((3 + 8) + 20) - 8)", "((8 + (3 + 20)) - 8)", "((3 + (8 + 20))...
((8 + 20) - (8 - 3))
bucket_a
((8 + 20) - (8 - 3))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.091042
21,001
{ "canonical_solution": "((8 + 20) - (8 - 3))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
22
[ 1, 25 ]
[ 3, 8, 8, 20 ]
{ "completion_tokens": 2262, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02220
Using the numbers [3, 8, 8, 20], write an arithmetic expression that equals 23. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
23
[ "(17 + 8)", "(((17 - 7) - 8) + 23)", "(((17 - 8) - 7) + 23)", "((17 - (7 + 8)) + 23)", "((17 + 23) - (7 + 8))", "((7 * 23) - (17 * 8))", "((17 - 8) + (23 - 7))", "(((17 - 7) + 23) - 8)", "(((17 + 23) - 7) - 8)", "((17 + (23 - 7)) - 8)", "((17 - 7) + (23 - 8))", "(((17 - 8) + 23) - 7)", "(((1...
(17 + 8)
bucket_a
(17 + 8)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.833213
21,001
{ "canonical_solution": "(17 + 8)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
17
[ 1, 25 ]
[ 17, 7, 8, 23 ]
{ "completion_tokens": 2308, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02224
Using the numbers [17, 7, 8, 23], write an arithmetic expression that equals 25. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
25
[ "(22 + 22)", "((1 * 22) + 22)", "((22 / 1) + 22)", "(22 + (1 * 22))", "(22 + (22 / 1))", "(1 * (22 + 22))", "((22 + 22) / 1)" ]
((22 + 22) * 1)
bucket_a
((22 + 22) * 1)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "((22 + 22) * 1)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
7
[ 1, 25 ]
[ 1, 22, 6, 22 ]
{ "completion_tokens": 3914, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02231
Using the numbers [1, 22, 6, 22], write an arithmetic expression that equals 44. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
44
[ "(16 + 6)", "22", "(((16 - 6) / 10) * 22)", "(22 / ((16 - 6) / 10))", "((10 / (16 - 6)) * 22)", "(22 / (10 / (16 - 6)))", "((6 / (16 - 10)) * 22)", "(22 / (6 / (16 - 10)))", "(((16 - 10) / 6) * 22)", "(22 / ((16 - 10) / 6))", "((16 / (6 + 10)) * 22)", "(22 / (16 / (6 + 10)))", "(((6 + 10) / ...
(6 + 16)
bucket_a
(6 + 16)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.295837
21,001
{ "canonical_solution": "(6 + 16)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
27
[ 1, 25 ]
[ 16, 6, 10, 22 ]
{ "completion_tokens": 2372, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02236
Using the numbers [16, 6, 10, 22], write an arithmetic expression that equals 22. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
22
[ "((7 - 3) + 25)", "(7 + (25 - 3))", "((7 + 25) - 3)", "(((14 - 3) - 7) + 25)", "((14 - (3 + 7)) + 25)", "(((14 - 7) - 3) + 25)", "((14 - 7) + (25 - 3))", "((14 + 25) - (3 + 7))", "(((14 - 3) + 25) - 7)", "((14 + (25 - 3)) - 7)", "(((14 + 25) - 3) - 7)", "((14 - 3) + (25 - 7))", "(14 + (25 - ...
(7 + (25 - 3))
bucket_a
(7 + (25 - 3))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "(7 + (25 - 3))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
18
[ 1, 25 ]
[ 3, 14, 7, 25 ]
{ "completion_tokens": 2649, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "parse_error": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02261
Using the numbers [3, 14, 7, 25], write an arithmetic expression that equals 29. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
29
[ "(((2 * 14) - 2) + 19)", "(((14 * 2) - 2) + 19)", "((14 * 2) + (19 - 2))", "((14 / 2) + (2 * 19))", "(((2 * 14) + 19) - 2)", "((2 * 14) + (19 - 2))", "(((14 * 2) + 19) - 2)" ]
((2 * 19) + (14 / 2))
bucket_a
((2 * 19) + (14 / 2))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "((2 * 19) + (14 / 2))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
7
[ 1, 25 ]
[ 2, 14, 2, 19 ]
{ "completion_tokens": 4576, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "parse_error": 1, "wrong_value": 5 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02264
Using the numbers [2, 14, 2, 19], write an arithmetic expression that equals 45. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
45
[ "(21 + 4)", "((1 * 21) + 4)", "((21 / 1) + 4)", "(21 + (1 * 4))", "(21 + (4 / 1))", "(1 * (21 + 4))", "((21 + 4) / 1)", "(((9 - 1) + 21) - 4)", "((9 + (21 - 1)) - 4)", "(((9 + 21) - 1) - 4)", "((9 + 21) - (1 + 4))", "((21 - 1) + (9 - 4))", "(21 + ((9 - 1) - 4))", "(21 + (9 - (1 + 4)))", ...
(4 + 21)
bucket_a
(4 + 21)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.091042
21,001
{ "canonical_solution": "(4 + 21)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
22
[ 1, 25 ]
[ 1, 9, 21, 4 ]
{ "completion_tokens": 1889, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02276
Using the numbers [1, 9, 21, 4], write an arithmetic expression that equals 25. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
25
[ "(((10 - 3) + 12) + 13)", "((10 + (12 - 3)) + 13)", "(((10 + 12) - 3) + 13)", "((10 + 12) + (13 - 3))", "((12 - 10) * (3 + 13))", "((12 - 3) + (10 + 13))", "(12 + ((10 - 3) + 13))", "(12 + (10 + (13 - 3)))", "(12 + ((10 + 13) - 3))", "((10 - 3) + (12 + 13))", "(10 + ((12 - 3) + 13))", "(10 + (...
(((13 + 10) - 3) + 12)
bucket_a
(((13 + 10) - 3) + 12)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "(((13 + 10) - 3) + 12)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 3, 10, 12, 13 ]
{ "completion_tokens": 3434, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 1, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02288
Using the numbers [3, 10, 12, 13], write an arithmetic expression that equals 32. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
32
[ "((8 * 6) + 5)", "((8 + 15) + (6 * 5))", "(15 + (8 + (6 * 5)))", "(8 + ((15 - 6) * 5))", "(8 + (15 + (6 * 5)))" ]
((8 + 15) + (5 * 6))
bucket_a
((8 + 15) + (5 * 6))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.609438
21,001
{ "canonical_solution": "((8 + 15) + (5 * 6))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
5
[ 1, 25 ]
[ 8, 15, 6, 5 ]
{ "completion_tokens": 4775, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "parse_error": 3, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02292
Using the numbers [8, 15, 6, 5], write an arithmetic expression that equals 53. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
53
[ "((7 * 8) - 22)", "(((13 + 22) + 7) - 8)", "((22 + (13 + 7)) - 8)", "((13 + (22 + 7)) - 8)", "((22 + 7) + (13 - 8))", "((13 + 7) + (22 - 8))", "(7 + ((13 + 22) - 8))", "(7 + (22 + (13 - 8)))", "(7 + (13 + (22 - 8)))", "((13 + 22) - (8 - 7))", "(22 + ((13 + 7) - 8))", "(22 + (7 + (13 - 8)))", ...
(((22 + 7) - 8) + 13)
bucket_a
(((22 + 7) - 8) + 13)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "(((22 + 7) - 8) + 13)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 13, 22, 7, 8 ]
{ "completion_tokens": 2832, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02306
Using the numbers [13, 22, 7, 8], write an arithmetic expression that equals 34. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
34
[ "((18 + 22) + 3)", "(22 + (18 + 3))", "(18 + (22 + 3))" ]
((18 + 3) + 22)
bucket_a
((18 + 3) + 22)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.098612
21,001
{ "canonical_solution": "((18 + 3) + 22)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
3
[ 1, 25 ]
[ 20, 18, 22, 3 ]
{ "completion_tokens": 3072, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02310
Using the numbers [20, 18, 22, 3], write an arithmetic expression that equals 43. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
43
[ "((1 + 10) + 12)", "(10 + (1 + 12))", "(1 + (10 + 12))", "((1 + 12) + 10)", "(12 + (1 + 10))", "(1 + (12 + 10))" ]
((12 + 1) + 10)
bucket_a
((12 + 1) + 10)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.791759
21,001
{ "canonical_solution": "((12 + 1) + 10)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
6
[ 1, 25 ]
[ 1, 10, 12, 10 ]
{ "completion_tokens": 4335, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02319
Using the numbers [1, 10, 12, 10], write an arithmetic expression that equals 23. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
23
[ "(((14 + 10) + 7) + 3)", "((10 + (14 + 7)) + 3)", "((14 + (10 + 7)) + 3)", "((10 + 7) + (14 + 3))", "((14 + 7) + (10 + 3))", "(7 + ((14 + 10) + 3))", "(7 + (10 + (14 + 3)))", "(7 + (14 + (10 + 3)))", "((14 + 10) + (7 + 3))", "(10 + ((14 + 7) + 3))", "(10 + (7 + (14 + 3)))", "(10 + (14 + (7 + 3...
(10 + ((3 + 7) + 14))
bucket_a
(10 + ((3 + 7) + 14))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(10 + ((3 + 7) + 14))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 14, 10, 7, 3 ]
{ "completion_tokens": 2959, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "parse_error": 3, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02321
Using the numbers [14, 10, 7, 3], write an arithmetic expression that equals 34. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
34
[ "(21 - (15 / 3))", "(23 - (21 / 3))", "(((21 + 15) - 23) + 3)", "((15 - (23 - 21)) + 3)", "((21 - (23 - 15)) + 3)", "((21 + 3) - (23 - 15))", "((15 + 3) - (23 - 21))", "(((21 + 15) + 3) - 23)", "((15 + (21 + 3)) - 23)", "((21 + (15 + 3)) - 23)", "((21 + 15) - (23 - 3))", "(15 + (3 - (23 - 21))...
(((21 + 3) + 15) - 23)
bucket_a
(((21 + 3) + 15) - 23)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.833213
21,001
{ "canonical_solution": "(((21 + 3) + 15) - 23)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
17
[ 1, 25 ]
[ 21, 15, 23, 3 ]
{ "completion_tokens": 2250, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "non_positive_intermediate": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02336
Using the numbers [21, 15, 23, 3], write an arithmetic expression that equals 16. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
16
[ "(9 + 7)", "((18 - 9) + 7)", "((18 + 7) - 9)", "(18 - (9 - 7))", "((18 + 9) - 11)", "(9 + (18 - 11))", "(18 - (11 - 9))", "((7 - (18 / 9)) + 11)", "(7 + (11 - (18 / 9)))", "(7 + (18 / (11 - 9)))", "((7 + 11) - (18 / 9))", "(18 - ((7 + 11) / 9))" ]
(7 + 9)
bucket_a
(7 + 9)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.484907
21,001
{ "canonical_solution": "(7 + 9)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_a...
4
12
[ 1, 25 ]
[ 18, 9, 7, 11 ]
{ "completion_tokens": 4131, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02337
Using the numbers [18, 9, 7, 11], write an arithmetic expression that equals 16. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
16
[ "((8 + 9) + 1)", "(9 + (8 + 1))", "(8 + (9 + 1))", "(((20 + 8) - 9) - 1)", "((8 + (20 - 9)) - 1)", "((20 - (9 - 8)) - 1)", "((20 - 1) - (9 - 8))", "((20 - 9) + (8 - 1))", "(((20 + 8) - 1) - 9)", "((8 + (20 - 1)) - 9)", "((20 + (8 - 1)) - 9)", "((20 + 8) - (9 + 1))", "(8 + ((20 - 9) - 1))", ...
(9 + (8 + 1))
bucket_a
(9 + (8 + 1))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "(9 + (8 + 1))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
18
[ 1, 25 ]
[ 20, 8, 9, 1 ]
{ "completion_tokens": 3469, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 4, "non_positive_intermediate": 1, "parse_error": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02341
Using the numbers [20, 8, 9, 1], write an arithmetic expression that equals 18. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
18
[ "((12 + 14) + 21)", "(14 + (12 + 21))", "(12 + (14 + 21))", "(((24 - 12) + 14) + 21)", "((24 + (14 - 12)) + 21)", "(((24 + 14) - 12) + 21)", "((24 + 14) + (21 - 12))", "((14 - 12) + (24 + 21))", "(14 + ((24 - 12) + 21))", "(14 + (24 + (21 - 12)))", "(14 + ((24 + 21) - 12))", "((24 - 12) + (14 ...
((21 + 14) + 12)
bucket_a
((21 + 14) + 12)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((21 + 14) + 12)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "r...
4
18
[ 1, 25 ]
[ 12, 24, 14, 21 ]
{ "completion_tokens": 2075, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02349
Using the numbers [12, 24, 14, 21], write an arithmetic expression that equals 47. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as a...
train
easy
47
[ "(((19 - 4) + 14) + 3)", "((19 + (14 - 4)) + 3)", "(((19 + 14) - 4) + 3)", "((19 + 14) - (4 - 3))", "((14 - 4) + (19 + 3))", "(14 + ((19 - 4) + 3))", "(14 + (19 - (4 - 3)))", "(14 + ((19 + 3) - 4))", "((19 - 4) + (14 + 3))", "(19 + ((14 - 4) + 3))", "(19 + (14 - (4 - 3)))", "(19 + ((14 + 3) - ...
((19 - 4) + (14 + 3))
bucket_a
((19 - 4) + (14 + 3))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((19 - 4) + (14 + 3))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
18
[ 1, 25 ]
[ 4, 19, 14, 3 ]
{ "completion_tokens": 4902, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_integer_intermediate": 1, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02361
Using the numbers [4, 19, 14, 3], write an arithmetic expression that equals 32. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
32
[ "((7 * 5) - 8)", "(((17 + 7) + 8) - 5)", "((7 + (17 + 8)) - 5)", "((17 + (7 + 8)) - 5)", "((7 + 8) + (17 - 5))", "((17 + 8) + (7 - 5))", "(8 + ((17 + 7) - 5))", "(8 + (7 + (17 - 5)))", "(8 + (17 + (7 - 5)))", "((17 + 7) + (8 - 5))", "(7 + ((17 + 8) - 5))", "(7 + (8 + (17 - 5)))", "(7 + (17 +...
((7 * 5) - 8)
bucket_a
((7 * 5) - 8)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "((7 * 5) - 8)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
16
[ 1, 25 ]
[ 17, 7, 8, 5 ]
{ "completion_tokens": 4394, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02365
Using the numbers [17, 7, 8, 5], write an arithmetic expression that equals 27. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
27
[ "((5 * (15 - 8)) - 2)", "((8 + 15) + (5 * 2))", "(15 + (8 + (5 * 2)))", "((8 - 5) + (15 * 2))", "(8 + (15 + (5 * 2)))", "(8 + ((15 * 2) - 5))", "((8 + (15 * 2)) - 5)" ]
(8 + ((5 * 2) + 15))
bucket_a
(8 + ((5 * 2) + 15))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.94591
21,001
{ "canonical_solution": "(8 + ((5 * 2) + 15))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
7
[ 1, 25 ]
[ 5, 8, 15, 2 ]
{ "completion_tokens": 5330, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "no_box": 1, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02370
Using the numbers [5, 8, 15, 2], write an arithmetic expression that equals 33. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
33
[ "(((1 + 18) + 13) - 3)", "((18 + (1 + 13)) - 3)", "((1 + (18 + 13)) - 3)", "((18 + 13) - (3 - 1))", "((1 + 13) + (18 - 3))", "(13 + ((1 + 18) - 3))", "(13 + (18 - (3 - 1)))", "(13 + (1 + (18 - 3)))", "((1 + 18) + (13 - 3))", "(18 + ((1 + 13) - 3))", "(18 + (13 - (3 - 1)))", "(18 + (1 + (13 - 3...
((13 - 3) + (18 + 1))
bucket_a
((13 - 3) + (18 + 1))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((13 - 3) + (18 + 1))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 1, 18, 13, 3 ]
{ "completion_tokens": 3677, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02374
Using the numbers [1, 18, 13, 3], write an arithmetic expression that equals 29. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
29
[ "(16 * 6)", "((16 * 6) * 1)", "((16 * 6) / 1)", "(6 * (16 * 1))", "(6 * (16 / 1))", "(16 * (6 * 1))", "(16 * (6 / 1))", "((16 * 1) * 6)", "((16 / 1) * 6)", "(1 * (16 * 6))", "(16 * (1 * 6))", "((6 * (16 + 1)) - 6)", "((6 * (16 - 1)) + 6)", "(((16 + 1) * 6) - 6)", "(6 + ((16 - 1) * 6))" ]
(16 * 6)
bucket_a
(16 * 6)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(16 * 6)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
15
[ 1, 25 ]
[ 16, 6, 1, 6 ]
{ "completion_tokens": 3285, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "wrong_value": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02386
Using the numbers [16, 6, 1, 6], write an arithmetic expression that equals 96. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
96
[ "(((19 + 13) - 7) + 23)", "((13 + (19 - 7)) + 23)", "((19 + (13 - 7)) + 23)", "((13 - 7) + (19 + 23))", "((19 - 7) + (13 + 23))", "(((19 + 13) + 23) - 7)", "((13 + (19 + 23)) - 7)", "((19 + (13 + 23)) - 7)", "((19 + 13) + (23 - 7))", "(13 + ((19 - 7) + 23))", "(13 + ((19 + 23) - 7))", "(13 + (...
(19 + ((13 + 23) - 7))
bucket_a
(19 + ((13 + 23) - 7))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "(19 + ((13 + 23) - 7))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 19, 13, 7, 23 ]
{ "completion_tokens": 4860, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 3, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02387
Using the numbers [19, 13, 7, 23], write an arithmetic expression that equals 48. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
48
[ "((5 + 9) + 25)", "(9 + (5 + 25))", "(5 + (9 + 25))", "((18 - (9 - 5)) + 25)", "(((5 + 18) - 9) + 25)", "((5 + (18 - 9)) + 25)", "((18 - 9) + (5 + 25))", "((5 + 18) + (25 - 9))", "(18 + (25 - (9 - 5)))", "(18 + ((5 + 25) - 9))", "(18 + (5 + (25 - 9)))", "((18 + 25) - (9 - 5))", "(((5 + 18) +...
(9 + (25 + 5))
bucket_a
(9 + (25 + 5))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "(9 + (25 + 5))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
18
[ 1, 25 ]
[ 5, 9, 18, 25 ]
{ "completion_tokens": 3789, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 2, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02405
Using the numbers [5, 9, 18, 25], write an arithmetic expression that equals 39. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
39
[ "(((11 + 25) + 3) + 21)", "((25 + (11 + 3)) + 21)", "((11 + (25 + 3)) + 21)", "((25 + 3) + (11 + 21))", "((11 + 3) + (25 + 21))", "(3 + ((11 + 25) + 21))", "(3 + (25 + (11 + 21)))", "(3 + (11 + (25 + 21)))", "((11 + 25) + (3 + 21))", "(25 + ((11 + 3) + 21))", "(25 + (3 + (11 + 21)))", "(25 + (...
((21 + 3) + (25 + 11))
bucket_a
((21 + 3) + (25 + 11))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((21 + 3) + (25 + 11))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 11, 25, 3, 21 ]
{ "completion_tokens": 3750, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "invalid_operands": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02418
Using the numbers [11, 25, 3, 21], write an arithmetic expression that equals 60. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
60
[ "(10 + (2 * 12))", "((2 + (2 * 10)) + 12)", "((2 * 10) + (2 + 12))", "(2 + ((2 * 10) + 12))", "(2 * ((10 / 2) + 12))" ]
(2 + (12 + (2 * 10)))
bucket_a
(2 + (12 + (2 * 10)))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.609438
21,001
{ "canonical_solution": "(2 + (12 + (2 * 10)))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
5
[ 1, 25 ]
[ 2, 2, 10, 12 ]
{ "completion_tokens": 5129, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02440
Using the numbers [2, 2, 10, 12], write an arithmetic expression that equals 34. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
34
[ "((22 + 17) - 9)", "(17 + (22 - 9))", "(22 + (17 - 9))", "(((16 + 22) - 17) + 9)", "((22 - (17 - 16)) + 9)", "((16 + (22 - 17)) + 9)", "((22 - 17) + (16 + 9))", "((22 + 9) - (17 - 16))", "(((16 + 22) + 9) - 17)", "((22 + (16 + 9)) - 17)", "((16 + (22 + 9)) - 17)", "((16 + 22) - (17 - 9))", "...
((17 + 22) - 9)
bucket_a
((17 + 22) - 9)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.890372
21,001
{ "canonical_solution": "((17 + 22) - 9)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
18
[ 1, 25 ]
[ 16, 22, 17, 9 ]
{ "completion_tokens": 3176, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 2, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02443
Using the numbers [16, 22, 17, 9], write an arithmetic expression that equals 30. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
30
[ "(19 - 5)", "(5 + 9)", "((19 - 5) * 1)", "((19 - 5) / 1)", "(19 - (5 * 1))", "(19 - (5 / 1))", "((19 * 1) - 5)", "((19 / 1) - 5)", "((5 + 9) * 1)", "((5 + 9) / 1)", "(9 + (5 * 1))", "(9 + (5 / 1))", "(5 + (9 * 1))", "(5 + (9 / 1))", "(((5 + 19) - 9) - 1)", "((19 - (9 - 5)) - 1)", "((...
(5 + 9)
bucket_a
(5 + 9)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.367296
21,001
{ "canonical_solution": "(5 + 9)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_a...
4
29
[ 1, 25 ]
[ 5, 19, 9, 1 ]
{ "completion_tokens": 1759, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.6, "reason_counts": { "correct": 6, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 6 }
cd_bucket_a_train_21001_02482
Using the numbers [5, 19, 9, 1], write an arithmetic expression that equals 14. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
14
[ "((20 - 15) + (3 * 12))", "((20 + (3 * 12)) - 15)", "(20 + ((3 * 12) - 15))" ]
(20 + ((12 * 3) - 15))
bucket_a
(20 + ((12 * 3) - 15))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.098612
21,001
{ "canonical_solution": "(20 + ((12 * 3) - 15))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
3
[ 1, 25 ]
[ 20, 15, 3, 12 ]
{ "completion_tokens": 6571, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "no_box": 1, "non_positive_intermediate": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02483
Using the numbers [20, 15, 3, 12], write an arithmetic expression that equals 41. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
41
[ "((24 + 16) - 5)", "(16 + (24 - 5))", "(24 + (16 - 5))", "((14 + 16) + 5)", "(16 + (14 + 5))", "(14 + (16 + 5))" ]
(16 + (14 + 5))
bucket_a
(16 + (14 + 5))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-1.791759
21,001
{ "canonical_solution": "(16 + (14 + 5))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "re...
4
6
[ 1, 25 ]
[ 24, 14, 16, 5 ]
{ "completion_tokens": 3651, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "parse_error": 3, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02488
Using the numbers [24, 14, 16, 5], write an arithmetic expression that equals 35. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
35
[ "(4 + 8)", "(4 * (11 - 8))", "((25 * 4) - (11 * 8))", "(4 * ((25 + 8) / 11))", "((25 - 11) - (8 / 4))", "(25 - (11 + (8 / 4)))", "((4 * (25 + 8)) / 11)", "((25 - (8 / 4)) - 11)" ]
((11 - 8) * 4)
bucket_a
((11 - 8) * 4)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.079442
21,001
{ "canonical_solution": "((11 - 8) * 4)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
8
[ 1, 25 ]
[ 11, 25, 4, 8 ]
{ "completion_tokens": 3090, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02495
Using the numbers [11, 25, 4, 8], write an arithmetic expression that equals 12. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
12
[ "(((25 - 15) - 1) + 24)", "((25 - (15 + 1)) + 24)", "(((25 - 1) - 15) + 24)", "((25 - 1) + (24 - 15))", "((25 + 24) - (15 + 1))", "(((25 - 15) + 24) - 1)", "((25 + (24 - 15)) - 1)", "(((25 + 24) - 15) - 1)", "((25 - 15) + (24 - 1))", "(25 + (24 - (15 + 1)))", "(25 + ((24 - 15) - 1))", "(25 + (...
((24 - (1 + 15)) + 25)
bucket_a
((24 - (1 + 15)) + 25)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((24 - (1 + 15)) + 25)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 15, 25, 1, 24 ]
{ "completion_tokens": 1586, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 1, "non_positive_intermediate": 1, "parse_error": 1, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02499
Using the numbers [15, 25, 1, 24], write an arithmetic expression that equals 33. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
33
[ "((17 * 2) - 10)", "(19 + (10 / 2))", "(((17 + 19) - 2) - 10)", "((19 + (17 - 2)) - 10)", "((17 + (19 - 2)) - 10)", "((19 - 2) + (17 - 10))", "((17 - 2) + (19 - 10))", "(((17 + 19) - 10) - 2)", "((19 + (17 - 10)) - 2)", "((17 + (19 - 10)) - 2)", "(2 * ((19 - 17) + 10))", "(2 * (19 - (17 - 10))...
(2 * (19 - (17 - 10)))
bucket_a
(2 * (19 - (17 - 10)))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.044522
21,001
{ "canonical_solution": "(2 * (19 - (17 - 10)))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
21
[ 1, 25 ]
[ 17, 19, 2, 10 ]
{ "completion_tokens": 2838, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "invalid_operands": 2, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02504
Using the numbers [17, 19, 2, 10], write an arithmetic expression that equals 24. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
24
[ "(((25 + 2) + 19) - 6)", "((2 + (25 + 19)) - 6)", "((25 + (2 + 19)) - 6)", "((2 + 19) + (25 - 6))", "((25 + 19) - (6 - 2))", "(19 + ((25 + 2) - 6))", "(19 + (2 + (25 - 6)))", "(19 + (25 - (6 - 2)))", "((25 + 2) + (19 - 6))", "(2 + ((25 + 19) - 6))", "(2 + (19 + (25 - 6)))", "(2 + (25 + (19 - 6...
((25 - (6 - 2)) + 19)
bucket_a
((25 - (6 - 2)) + 19)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.70805
21,001
{ "canonical_solution": "((25 - (6 - 2)) + 19)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
15
[ 1, 25 ]
[ 25, 2, 19, 6 ]
{ "completion_tokens": 4780, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 3, "parse_error": 2, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02529
Using the numbers [25, 2, 19, 6], write an arithmetic expression that equals 40. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
40
[ "21", "((21 + 8) - 8)", "((21 * 8) / 8)", "((21 - 8) + 8)", "(8 + (21 - 8))", "(21 * (8 / 8))", "(21 / (8 / 8))", "(20 + (8 / 8))" ]
(21 / (8 / 8))
bucket_a
(21 / (8 / 8))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.079442
21,001
{ "canonical_solution": "(21 / (8 / 8))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
8
[ 1, 25 ]
[ 21, 20, 8, 8 ]
{ "completion_tokens": 2017, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "non_positive_intermediate": 2, "parse_error": 3 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02542
Using the numbers [21, 20, 8, 8], write an arithmetic expression that equals 21. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
21
[ "((3 * (13 + 12)) + 13)", "(13 + (3 * (12 + 13)))" ]
(13 + ((12 + 13) * 3))
bucket_a
(13 + ((12 + 13) * 3))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-0.693147
21,001
{ "canonical_solution": "(13 + ((12 + 13) * 3))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
2
[ 1, 25 ]
[ 13, 3, 12, 13 ]
{ "completion_tokens": 7144, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "no_box": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02543
Using the numbers [13, 3, 12, 13], write an arithmetic expression that equals 88. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
88
[ "(24 - (22 - 19))", "((19 + 24) - 22)", "(19 + (24 - 22))", "(((19 + 4) + 22) - 24)", "((4 + (19 + 22)) - 24)", "((19 + (4 + 22)) - 24)", "((4 + 22) - (24 - 19))", "((19 + 22) - (24 - 4))", "(22 - (24 - (19 + 4)))", "(22 - ((24 - 19) - 4))", "(22 - ((24 - 4) - 19))", "((19 + 4) - (24 - 22))", ...
((19 + 22) - (24 - 4))
bucket_a
((19 + 22) - (24 - 4))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.944439
21,001
{ "canonical_solution": "((19 + 22) - (24 - 4))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
19
[ 1, 25 ]
[ 19, 4, 22, 24 ]
{ "completion_tokens": 2564, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "non_positive_intermediate": 3, "parse_error": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02545
Using the numbers [19, 4, 22, 24], write an arithmetic expression that equals 21. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an...
train
easy
21
[ "(7 + 23)", "((7 + 23) * 1)", "((7 + 23) / 1)", "(23 + (7 * 1))", "(23 + (7 / 1))", "(7 + (23 * 1))", "(7 + (23 / 1))", "(((15 - 7) + 23) - 1)", "(((15 + 23) - 7) - 1)", "((15 + (23 - 7)) - 1)", "((23 - 7) + (15 - 1))", "((15 + 23) - (7 + 1))", "(23 + ((15 - 7) - 1))", "(23 + ((15 - 1) - 7...
(7 + 23)
bucket_a
(7 + 23)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.091042
21,001
{ "canonical_solution": "(7 + 23)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_...
4
22
[ 1, 25 ]
[ 15, 7, 23, 1 ]
{ "completion_tokens": 3161, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "parse_error": 1, "wrong_value": 3 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02548
Using the numbers [15, 7, 23, 1], write an arithmetic expression that equals 30. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
30
[ "(15 * (8 - 6))", "(((17 + 6) + 15) - 8)", "((6 + (17 + 15)) - 8)", "((17 + (6 + 15)) - 8)", "((6 + 15) + (17 - 8))", "((17 + 15) - (8 - 6))", "(15 + ((17 + 6) - 8))", "(15 + (6 + (17 - 8)))", "(15 + (17 - (8 - 6)))", "((17 + 6) + (15 - 8))", "(6 + ((17 + 15) - 8))", "(6 + (15 + (17 - 8)))", ...
((15 + 6) + (17 - 8))
bucket_a
((15 + 6) + (17 - 8))
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.772589
21,001
{ "canonical_solution": "((15 + 6) + (17 - 8))", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
16
[ 1, 25 ]
[ 17, 6, 15, 8 ]
{ "completion_tokens": 6554, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 2, "non_positive_intermediate": 2, "parse_error": 1, "wrong_value": 1 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02549
Using the numbers [17, 6, 15, 8], write an arithmetic expression that equals 30. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
30
[ "(8 * (2 + 1))", "((8 * 2) + 8)", "(8 + (2 * 8))", "((2 + 1) * 8)", "(((8 * 2) * 1) + 8)", "(((8 * 2) / 1) + 8)", "((2 * (8 * 1)) + 8)", "((2 * (8 / 1)) + 8)", "((8 * (2 * 1)) + 8)", "((8 * (2 / 1)) + 8)", "(((8 / 2) - 1) * 8)", "((8 * 1) + (2 * 8))", "((8 / 1) + (2 * 8))", "(1 * ((8 * 2) ...
((8 * 2) + 8)
bucket_a
((8 * 2) + 8)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.258097
21,001
{ "canonical_solution": "((8 * 2) + 8)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "requ...
4
26
[ 1, 25 ]
[ 8, 2, 1, 8 ]
{ "completion_tokens": 3310, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "invalid_operands": 1, "non_positive_intermediate": 1, "wrong_value": 4 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02553
Using the numbers [8, 2, 1, 8], write an arithmetic expression that equals 24. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ex...
train
easy
24
[ "(24 - (1 + 6))", "((24 - 1) - 6)", "((24 - 6) - 1)", "(24 - ((2 - 1) + 6))", "(24 - ((2 + 6) - 1))", "(24 - (2 + (6 - 1)))", "((24 - 2) - (6 - 1))", "((6 - 1) + (24 / 2))", "((1 + 24) - (2 + 6))", "((24 - (2 - 1)) - 6)", "((1 + (24 - 2)) - 6)", "(((1 + 24) - 2) - 6)", "(6 + ((24 / 2) - 1))"...
((24 - 1) - 6)
bucket_a
((24 - 1) - 6)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-3.044522
21,001
{ "canonical_solution": "((24 - 1) - 6)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "req...
4
21
[ 1, 25 ]
[ 2, 1, 6, 24 ]
{ "completion_tokens": 2460, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.4, "reason_counts": { "correct": 4, "parse_error": 4, "wrong_value": 2 }, "rollouts": 10, "success_count": 4 }
cd_bucket_a_train_21001_02562
Using the numbers [2, 1, 6, 24], write an arithmetic expression that equals 17. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an e...
train
easy
17
[ "(((9 + 16) + 12) - 5)", "((16 + (9 + 12)) - 5)", "((9 + (16 + 12)) - 5)", "((16 + 12) + (9 - 5))", "((9 + 12) + (16 - 5))", "(12 + ((9 + 16) - 5))", "(12 + (16 + (9 - 5)))", "(12 + (9 + (16 - 5)))", "((9 + 16) + (12 - 5))", "(16 + ((9 + 12) - 5))", "(16 + (12 + (9 - 5)))", "(16 + (9 + (12 - 5...
(((12 + 16) + 9) - 5)
bucket_a
(((12 + 16) + 9) - 5)
{ "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, "required_any_ops": [], "required_ops": [], "solution_operands_max": 4, "solution_...
bucket_a
-2.944439
21,001
{ "canonical_solution": "(((12 + 16) + 9) - 5)", "config": { "allowed_ops": [ "+", "-", "*", "/" ], "enumeration_cap": 1000, "max_generation_attempts": 20000, "max_solutions": 500, "min_solutions": 1, "n_operands": 4, "operand_max": 25, "operand_min": 1, ...
4
19
[ 1, 25 ]
[ 9, 16, 12, 5 ]
{ "completion_tokens": 5203, "model": "Qwen/Qwen3-0.6B", "observed_success_fraction": 0.5, "reason_counts": { "correct": 5, "no_box": 1, "parse_error": 2, "wrong_value": 2 }, "rollouts": 10, "success_count": 5 }
cd_bucket_a_train_21001_02565
Using the numbers [9, 16, 12, 5], write an arithmetic expression that equals 32. You may use +, -, *, / and parentheses. Each number may be used at most once; you do not need to use every number. Intermediate values must be positive integers. Keep any reasoning brief. Write your final answer inside \boxed{...} as an ...
train
easy
32