CooperBench/qwen35-9b-reasoning-share-coop-random-50

What this is

Cooperative two-agent coding dataset: 50 task pairs across 15 repos (random-50 subset), generated with mini_swe_agent on Qwen/Qwen3.5-9B in coop setting using a reasoning-share prompt variant — agents share their internal reasoning and analysis with each other before and during implementation, giving each agent visibility into the other's thought process to improve integration. All 50 pairs were successfully evaluated.

At a glance

Field	Value
Model	Qwen/Qwen3.5-9B
Agent	mini_swe_agent (reasoning-share prompt)
Setting	coop
Subset	random-50 (50 completed)
Repos	15
Pairs evaluated	50
Both-pass	8.0% (4/50)
Per-feature pass	22.0% (22/100)
Merge clean rate	62.0% (31/50)
Merge conflict rate	38.0% (19/50)
Total tokens	~27.8M (in+out, from traj files)
Owner	Arya Prabhudesai
Date	2026-06-02

How it was generated

cooperbench run --setting coop -a mini_swe_agent -c 30 qwen35-9b-reasoning-share-coop-random-50

Model served via vLLM OpenAI-compatible endpoint (openai/Qwen/Qwen3.5-9B). Reasoning-share variant: agents explicitly share their reasoning and analysis steps with each other to give insight into how each is approaching their feature.

File layout

• index.csv — one row per task pair; HF Dataset Viewer entry point
• qwen35-9b-reasoning-share-coop-random-50/coop/<repo>/<task_id>/<features>/ — raw per-pair artifacts: result.json, eval.json, agent1_traj.json, agent2_traj.json, agent{1,2}.patch, conversation.json

log_dir column in index.csv points to the per-pair subdirectory.

Schema highlights for mid-training

Filter on: both_passed=true, model, agent_framework.

metadata JSON carries per-agent statuses, steps, merge outcome, per-feature pass — use json.loads(row["metadata"]) without following the pointer.

Note: total_tokens is 0 for this run — token counts are in agent_full_traj.json under messages[*].extra.response.usage (~27.8M total in+out).

Caveats

• random-50 subset — all 50 tasks completed
• 11.0% LimitsExceeded exits (11/100 agent slots); 3.0% Error (3/100)
• 38.0% merge conflict rate (19/50)
• Token fields in result.json are 0; aggregate from agent_full_traj.json if needed

Citation

@dataset{qwen35_9b_reasoning_share_coop_random_50,
  title  = {qwen35-9b-reasoning-share-coop-random-50},
  author = {Arya Prabhudesai},
  year   = {2026},
  url    = {https://huggingface.co/datasets/CooperBench/qwen35-9b-reasoning-share-coop-random-50}
}