---
license: apache-2.0
pipeline_tag: text-generation
---

# ROME-30B-A3B (Coming Soon)

<a href="https://arxiv.org/pdf/2512.24873" target="_blank">
  🔗 <strong>Technical Report</strong><br/>
  <img alt="Paper" src="https://img.shields.io/badge/Paper-arXiv%3A2512.24873-red"/>
</a>


---

## 📢 Note: Coming Soon!

**ROME (ROME is Obviously an Agentic ModEl)** will be officially released soon. 
The project is currently under final review and preparation. Model weights will be made publicly available shortly. Stay tuned!

<img src="https://rlhf.oss-cn-hangzhou.aliyuncs.com/iFLOW-ROME/performance.png" width="600"/>


---



## Highlights

**ROME** is an open-source **agentic model** incubated within the **ALE (Agentic Learning Ecosystem)**.  

Rather than scaling performance purely by increasing parameter count, ROME achieves parameter-scale–crossing agentic performance through full-stack infrastructure and RL algorithmic optimization.

<img src="https://rlhf.oss-cn-hangzhou.aliyuncs.com/iFLOW-ROME/ALE.PNG" width="600"/>


### 🔧 ALE Full-Stack Infrastructure
- **ROLL** – Large-scale reinforcement learning optimization engine  

- **ROCK** – Secure sandbox and environment orchestration for agent execution  

- **iFlow CLI** – Unified agent framework and developer interface  

  

### 🧠 IPA Policy Optimization Algorithm
- Introduces **Interaction-Perceptive Agentic Policy Optimization (IPA)**  
- Performs credit assignment at the level of **Semantic Interaction Chunks**  
- Significantly improves **training stability** and **success rates** on **long-horizon tasks**



### 🚀 Strong Agentic Performance
- Despite being a **mid-sized model** (30B MoE with 3B active parameters), ROME outperforms same-scale models on standard agent benchmarks:
  - **Terminal-Bench 2.0**: 24.72%
  - **SWE-bench Verified**: 57.40%
  
- Performance is competitive with, and in some cases comparable to, models exceeding **100B parameters**

  

### 🔒 Production-Grade Safety
- Designed for autonomous agent execution in real environments  
- Rigorously aligned and red-teamed against risks such as:
  - Unauthorized access
  - Illegal or unsafe tool invocation
- Built with **deployment-grade safety guarantees** in mind

---



## Performance (Preview)

### Terminal-Based Benchmarks

| **Model**                    | **Terminal-Bench 2.0** | **SWE-bench Verified** |
| ---------------------------- | ---------------------- | ---------------------- |
| Qwen3-Coder-30B-A3B-Instruct | 13.48%                 | 46.33%                 |
| **ROME-30B-A3B**             | **24.72%**             | **57.40%**             |
| GPT-OSS-120B                 | 21.12%                 | 43.93%                 |
| GLM-4.5 Air (106B)           | 17.30%                 | 56.20%                 |

> See the technical report for full experimental details.

---

## Best Practices

*(Code examples and usage guidelines will be added after the model release.)*

---



## Citation

If you find our work useful, please consider citing:

```bibtex
@article{rome2025ale,
  title={Let It Flow: Agentic Crafting on Rock and Roll - Building the ROME Model within an Open Agentic Learning Ecosystem},
  author={Wang, Weixun and Xu, XiaoXiao and An, Wanhe and Dai, Fangwen and others},
  journal={arXiv preprint arXiv:2512.24873},
  year={2025}
}