Modelpublic

cocoa-org/Mocha-Coder-32B

sourceHugging Facemitupdated 4mo agoView on Hugging Face

4likes32downloads

Model Card

<h1 style=" font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Helvetica,Arial,sans-serif; font-size:48px; font-weight:700; line-height:1.25; text-align:center; margin:0 0 24px;"> Mocha-Coder-32B </h1>

<p style="text-align:center; margin:0 0 8px; font-size:16px;"> <a href="https://junliwang.tech/">Junli Wang</a><sup></sup>   <a href="https://blankcheng.github.io/">Zhoujun Cheng</a><sup>†</sup>   <a href="https://yuxuan-zhang-dexter.github.io/">Yuxuan Zhang</a><sup>*</sup>   <a href="https://ber666.github.io/">Shibo Hao</a>   <a href="https://yaotang23.github.io/">Yao Tang</a>   <br> <a href="https://zhiting.ucsd.edu/">Zhiting Hu</a>   <a href="https://prithvirajva.com/">Prithviraj Ammanabrolu</a>   <a href="https://haozhang.ai/">Hao Zhang</a><sup>†</sup> </p>

<p style="text-align:center; margin:0 0 24px; font-size:14px; color:#555;"> University of California, San Diego  ·  <sup>*</sup>Equal Contribution  ·  <sup>†</sup>Corresponding Author </p>

<a href="https://github.com/cocoa-org/NanoRollout" style=" display:inline-block; padding:8px 24px; background:#2b2b2b; color:#ffffff; border-radius:36px; text-decoration:none; font-weight:600; font-size:16px;"> 🧑‍💻 NanoRollout Code </a>

<a href="https://huggingface.co/ZeonLap/Mocha-Coder-32B" style=" display:inline-block; padding:8px 24px; background:#2b2b2b; color:#ffffff; border-radius:36px; text-decoration:none; font-weight:600; font-size:16px;"> 🤗 Mocha-Coder-32B Model </a>

Introduction

Mocha-Coder-32B 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 is the state-of-the-art among open-data models and is competitive with much larger open-source models on agentic SWE benchmarks. </div>

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	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

</div>

SWE-Bench Pro

Model	Max Iteration	SWE-Bench Pro (Pass@1)
Qwen3-Coder-480B-A35B-Instruct	250	38.7
Mocha-Coder-32B	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

</div>

Terminal-Bench 2.0

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

</div>

Training Data

Mocha-Coder-32B 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	NanoRollout
SWE-Rebench	Qwen3-Coder-Next	mini-swe-agent	11.5	NanoRollout
SWE-Smith	Qwen3-Coder-480B-A35B	mini-swe-agent	12.8	NanoRollout
SWE-Smith	Qwen3-Coder-Next	mini-swe-agent	9.1	NanoRollout
SETA	Kimi-K2.5 / DeepSeek-V3.2	Terminus-2 JSON	14.0	NanoRollout

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

Running as an Agent

Mocha-Coder-32B 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 (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 or NanoRollout in your research, please cite NanoRollout:

bibtex

@misc{nanorollout,
  title  = {NanoRollout: A Lightweight Infra for Digital Agent Rollout at Scale},
  author = {Wang, Junli and Cheng, Zhoujun and Zhang, Yuxuan and Hao, Shibo
            and Tang, Yao and Hu, Zhiting and Ammanabrolu, Prithviraj
            and Zhang, Hao},
  year   = {2026},
  howpublished = {\url{https://github.com/cocoa-org/NanoRollout}},
}

</div>