base_model:
- Qwen/Qwen2.5-Coder-3B
license: mit
pipeline_tag: text-generation
library_name: transformers
Mixture-Code-Qwen2.5-Coder-3B
🌐 Project Page | 💻 Code | 📄 Paper
We introduce RecursiveMAS, a multi-agent framework that scales agent collaboration through latent-space recursion. RecursiveMAS treats a multi-agent system as a unified recursive computation, where heterogeneous agents iteratively exchange, refine, and evolve their latent states across recursion rounds. In the Mixture-Style setting, the Code Specialist Agent focuses on code generation and programming-related tasks, while collaborating with other domain-specialized agents through RecursiveLink modules for final response generation.
Model Details
| Item | Description |
|---|---|
| Model | Mixture-Code-Qwen2.5-Coder-3B |
| Collaboration Style | Mixture-Style |
| Agent Role | Code Specialist Agent |
| Base Model | Qwen2.5-Coder-3B |
⚠️ Note: This checkpoint is a role-specific agent in RecursiveMAS, rather than a standalone model intended for plain-text generation.
Sample Usage
To use this model as part of the RecursiveMAS system, you can load the entire system using the provided loader from the official repository:
from system_loader import load_mas_system
# Load the Mixture-Style MAS pipeline
mas = load_mas_system(
style="mixture",
device="cuda",
trust_remote_code=True,
)
# Access the Code Specialist agent
code_agent = mas.agents["code"].model
For detailed usage instructions and running on downstream tasks, please refer to the GitHub repository.
Model Collections for RecursiveMAS
| Style | Model Collection |
|---|---|
| Sequential-Style | 🤗 HuggingFace |
| Mixture-Style | 🤗 HuggingFace |
| Distillation-Style | 🤗 HuggingFace |
| Deliberation-Style | 🤗 HuggingFace |
Experiment Results
Citation
@misc{recursivemas,
title={Recursive Multi-Agent Systems},
author={Xiyuan Yang and Jiaru Zou and Rui Pan and Ruizhong Qiu and Pan Lu and Shizhe Diao and Jindong Jiang and Hanghang Tong and Tong Zhang and Markus J. Buehler and Jingrui He and James Zou},
year={2026},
eprint={2604.25917},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2604.25917},
}