Instructions to use JetBrains/Mellum2-12B-A2.5B-Thinking with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use JetBrains/Mellum2-12B-A2.5B-Thinking with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="JetBrains/Mellum2-12B-A2.5B-Thinking")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("JetBrains/Mellum2-12B-A2.5B-Thinking")
model = AutoModelForCausalLM.from_pretrained("JetBrains/Mellum2-12B-A2.5B-Thinking", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use JetBrains/Mellum2-12B-A2.5B-Thinking with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "JetBrains/Mellum2-12B-A2.5B-Thinking"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JetBrains/Mellum2-12B-A2.5B-Thinking",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/JetBrains/Mellum2-12B-A2.5B-Thinking

SGLang

How to use JetBrains/Mellum2-12B-A2.5B-Thinking with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "JetBrains/Mellum2-12B-A2.5B-Thinking" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JetBrains/Mellum2-12B-A2.5B-Thinking",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "JetBrains/Mellum2-12B-A2.5B-Thinking" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JetBrains/Mellum2-12B-A2.5B-Thinking",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use JetBrains/Mellum2-12B-A2.5B-Thinking with Docker Model Runner:
```
docker model run hf.co/JetBrains/Mellum2-12B-A2.5B-Thinking
```

pavlichenko commited on May 29

Commit

e6f65c2

verified ·

1 Parent(s): e3d2907

Update README.md

Browse files

Files changed (1) hide show

README.md +7 -7

README.md CHANGED Viewed

@@ -4,7 +4,7 @@ language:
 - en
 pipeline_tag: text-generation
 model-index:
-- name: Mellum 2 Thinking
   results:
   - task:
       type: text-generation
@@ -143,12 +143,12 @@ license: apache-2.0
 <img alt="Mellum" src="mellum-logo-dark.svg" width="320">
-# Mellum 2 Thinking
 > [!Note]
 > Use this model when you want explicit chain-of-thought before the final answer — complex debugging, multi-step planning, agentic workflows, and math- or reasoning-heavy tasks. For direct, low-latency answers without reasoning traces, use [Instruct](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Instruct) instead.
-## Mellum 2 Thinking Highlights
 Mellum 2 Thinking is a post-trained reasoning-augmented assistant model trained by JetBrains.
@@ -156,7 +156,7 @@ The model uses a Mixture-of-Experts architecture with 64 experts and activates 8
 It is produced from [`Mellum2-12B-A2.5B-Base`](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Base) by supervised fine-tuning (loss computed only on the final assistant turn) followed by reinforcement learning with verifiable rewards (RLVR) on a harder data mix that includes a long-form math subset. The model emits its reasoning inside `<think>...</think>` blocks before the final answer.
-## Mellum 2 Model Family
 This repository contains one checkpoint from the Mellum 2 family.
@@ -171,7 +171,7 @@ This repository contains one checkpoint from the Mellum 2 family.
 ## Model Overview
-**Mellum 2 Thinking** has the following features:
 - Number of Layers: 28
 - Hidden Size: 2304
@@ -232,7 +232,7 @@ print("Chat response:", chat_response)
 Post-training evaluation for the thinking/reasoning variants. All values are percentages; higher is better except HarmBench, where lower is better. All values self-reported by JetBrains.
-| Benchmark          | Mellum 2 Thinking SFT | Mellum 2 Thinking | Qwen3.5 (4B) | Qwen3.5 (9B) | OLMo-3 (7B) | Ministral 3 (14B) |
 | :----------------- | --------------------: | ----------------: | -----------: | -----------: | ----------: | ----------------: |
 | **Coding**         |                       |                   |              |              |             |                   |
 | LiveCodeBench v6   |                  75.1 |              69.9 |         59.4 |         68.3 |        59.8 |              42.7 |
@@ -260,7 +260,7 @@ Notes:
 - **JetBrains pairwise** is win rate against `Qwen2.5-7B-Instruct` on an internal benchmark.
 - `—` indicates the model lacks native tool calling (OLMo-3-7B-Thinking).
-For more details, see the [Mellum 2 Technical Report](TODO).
 ## License

 - en
 pipeline_tag: text-generation
 model-index:
+- name: Mellum2 Thinking
   results:
   - task:
       type: text-generation
 <img alt="Mellum" src="mellum-logo-dark.svg" width="320">
+# Mellum2 Thinking
 > [!Note]
 > Use this model when you want explicit chain-of-thought before the final answer — complex debugging, multi-step planning, agentic workflows, and math- or reasoning-heavy tasks. For direct, low-latency answers without reasoning traces, use [Instruct](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Instruct) instead.
+## Mellum2 Thinking Highlights
 Mellum 2 Thinking is a post-trained reasoning-augmented assistant model trained by JetBrains.
 It is produced from [`Mellum2-12B-A2.5B-Base`](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Base) by supervised fine-tuning (loss computed only on the final assistant turn) followed by reinforcement learning with verifiable rewards (RLVR) on a harder data mix that includes a long-form math subset. The model emits its reasoning inside `<think>...</think>` blocks before the final answer.
+## Mellum2 Model Family
 This repository contains one checkpoint from the Mellum 2 family.
 ## Model Overview
+**Mellum2 Thinking** has the following features:
 - Number of Layers: 28
 - Hidden Size: 2304
 Post-training evaluation for the thinking/reasoning variants. All values are percentages; higher is better except HarmBench, where lower is better. All values self-reported by JetBrains.
+| Benchmark          | Mellum2 Thinking SFT | Mellum2 Thinking | Qwen3.5 (4B) | Qwen3.5 (9B) | OLMo-3 (7B) | Ministral 3 (14B) |
 | :----------------- | --------------------: | ----------------: | -----------: | -----------: | ----------: | ----------------: |
 | **Coding**         |                       |                   |              |              |             |                   |
 | LiveCodeBench v6   |                  75.1 |              69.9 |         59.4 |         68.3 |        59.8 |              42.7 |
 - **JetBrains pairwise** is win rate against `Qwen2.5-7B-Instruct` on an internal benchmark.
 - `—` indicates the model lacks native tool calling (OLMo-3-7B-Thinking).
+For more details, see the [Mellum2 Technical Report](TODO).
 ## License