Instructions to use ZeonLap/Mocha-Coder-32B-Preview with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ZeonLap/Mocha-Coder-32B-Preview with Transformers:

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

pipe = pipeline("text-generation", model="ZeonLap/Mocha-Coder-32B-Preview")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ZeonLap/Mocha-Coder-32B-Preview")
model = AutoModelForCausalLM.from_pretrained("ZeonLap/Mocha-Coder-32B-Preview")
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

vLLM

How to use ZeonLap/Mocha-Coder-32B-Preview with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ZeonLap/Mocha-Coder-32B-Preview"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ZeonLap/Mocha-Coder-32B-Preview",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ZeonLap/Mocha-Coder-32B-Preview

SGLang

How to use ZeonLap/Mocha-Coder-32B-Preview 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 "ZeonLap/Mocha-Coder-32B-Preview" \
    --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": "ZeonLap/Mocha-Coder-32B-Preview",
		"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 "ZeonLap/Mocha-Coder-32B-Preview" \
        --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": "ZeonLap/Mocha-Coder-32B-Preview",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ZeonLap/Mocha-Coder-32B-Preview with Docker Model Runner:
```
docker model run hf.co/ZeonLap/Mocha-Coder-32B-Preview
```

Mocha-Coder-32B-Preview

Junli Wang^* Zhoujun Cheng^*† Yuxuan Zhang^* Shibo Hao
Zhiting Hu Prithviraj Ammanabrolu Hao Zhang^†

University of California, San Diego · ^*Equal Contribution · ^†Corresponding Author

🧑‍💻 NanoRollout Code

🤗 Mocha-Preview Model

📒 Wandb Log

Introduction

Mocha-Coder-32B-Preview is a strong open-data coding agent built on top of Qwen2.5-Coder-32B-Instruct. It is trained entirely through distillation on a 300K+ trajectory mixture sampled with our lightweight agent-rollout infrastructure, NanoRollout, with no reinforcement learning. The full training signal comes from frontier open-source teacher models (Qwen3-Coder-480B-A35B, Kimi-K2.5, Qwen3-Coder-Next, DeepSeek-V3.2) generating trajectories across multiple agent harnesses (OpenHands, mini-swe-agent, Terminus-2 JSON) on SWE-Rebench, SWE-Smith, and SETA.

The result is a simple but strong baseline coding agent: at the ≤32B scale, Mocha-Coder-32B-Preview is the state-of-the-art among open-data models and is competitive with much larger open-source models on agentic SWE benchmarks.

Key Features

Strong agentic SWE performance: 62.6 Pass@1 on SWE-Bench Verified, 35.3 on SWE-Bench Pro, 23.6 on Terminal-Bench 2.0, competitive with Qwen3-Coder-480B-A35B-Instruct.
Multi-harness training: Trajectories cover OpenHands, mini-swe-agent, and Terminus-2 JSON, mitigating harness-specific overfitting.
Open data: Distilled from a fully released 300K+ trajectory mixture (ZeonLap/Mocha-trajectories).

Performance

SWE-Bench Verified

Model	Max Iteration	SWE-Bench Verified (Pass@1)
Qwen3-Coder-480B-A35B-Instruct	100	67.0
Mocha-Coder-32B-Preview	100	62.6
SWE-Master-32B-RL	150	61.4
Kimi-Dev-72B	Agentless, TTS@40	60.4
CoderForge-Preview-32B	100	59.4
GLM-4.7-Flash	100	59.2
daVinci-Dev-72B	100	58.5
daVinci-Dev-32B	100	56.1
SERA-32B	100	54.2
Qwen3-Coder-30B-A3B-Instruct	100	51.6
Qwen2.5-Coder-32B-Instruct (Base)	100	6.2

SWE-Bench Pro

Model	Max Iteration	SWE-Bench Pro (Pass@1)
Qwen3-Coder-480B-A35B-Instruct	250	38.7
Mocha-Coder-32B-Preview	250	35.3
Gemini-3-flash	250	34.6
Kimi-K2-Instruct	250	27.7
DeepSeek-V3.2	250	15.6
Qwen2.5-Coder-32B-Instruct (Base)	250	0.0

Terminal-Bench 2.0

Model	Terminal-Bench 2.0
Qwen3-Coder-480B-A35B-Instruct	23.9
Mocha-Coder-32B-Preview	23.6
Qwen3-Coder-30B-A3B-Instruct	13.5
Qwen2.5-Coder-32B-Instruct (Base)	3.4

Training Data

Mocha-Coder-32B-Preview is trained on a 300K+ trajectory distillation mixture, drawn from previously released distillation sets (120K) and trajectories newly generated with NanoRollout (~180K).

Dataset	Teacher Model	Harness	# Trajectories (K)	Source
SWE-Rebench	Qwen3-Coder-480B-A35B	OpenHands	32.2	Nebius
SWE-Smith	Qwen3-Coder-480B-A35B	OpenHands	89.5	CoderForge
SWE-Rebench	Kimi-K2.5	mini-swe-agent	83.6	This release
SWE-Rebench	Qwen3-Coder-Next	mini-swe-agent	11.5	This release
SWE-Smith	Qwen3-Coder-480B-A35B	mini-swe-agent	12.8	This release
SWE-Smith	Qwen3-Coder-Next	mini-swe-agent	9.1	This release
SETA	Kimi-K2.5 / DeepSeek-V3.2	Terminus-2 JSON	14.0	This release

The full mixture is released at ZeonLap/Mocha-trajectories.

Running as an Agent

Mocha-Coder-32B-Preview is trained as an agent and is most useful when paired with a coding-agent harness. We have validated it with:

mini-swe-agent — minimal SWE agent loop, recommended for SWE-Bench Verified / Pro evaluation.
OpenHands — full-featured SWE harness; the model was trained on OpenHands trajectories.
Terminus-2 JSON — for Terminal-Bench 2.0 style shell tasks.

Point each harness's model endpoint at the vLLM server above. For SWE-Bench Verified we report numbers at a 100-iteration budget; for SWE-Bench Pro at 250 iterations.

License

Mocha-Coder-32B-Preview (model weights, training trajectories, and code) is released under the MIT License (see LICENSE) for research, educational, and commercial use.

Citation

If you use Mocha-Coder-32B-Preview or NanoRollout in your research, please cite:

@misc{wang2026mocha,
  title  = {Mocha-Coder-32B-Preview: Scaling Open-Data Coding Agents with NanoRollout},
  author = {Wang, Junli and Cheng, Zhoujun and Zhang, Yuxuan and Hao, Shibo
            and Hu, Zhiting and Ammanabrolu, Prithviraj and Zhang, Hao},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/ZeonLap/Mocha-Coder-32B-Preview}},
}

Downloads last month: 11

Safetensors

Model size

33B params

Tensor type

BF16

Model tree for ZeonLap/Mocha-Coder-32B-Preview

Base model

Qwen/Qwen2.5-32B

Finetuned

Qwen/Qwen2.5-Coder-32B

Finetuned

Qwen/Qwen2.5-Coder-32B-Instruct

Finetuned

(127)

this model