# Model Introduction
We introduce Agent Foundation Models (AFMs), a new family built on Qwen that natively perform end-to-end, multi-turn, multi-tool problem solving—without external frameworks or manual prompting. Built on the Chain-of-Agents (CoA) paradigm, each AFM dynamically activates specialized tool and role-playing agents inside a single forward pass, emulating the cooperative reasoning of a full multi-agent system. To train these models, we distilled high-performing multi-agent trajectories into agentic supervised-fine-tuning data and further optimized performance with agentic reinforcement learning on verifiable tasks. AFMs set new state-of-the-art results on benchmarks for both web and code agents, and we release all model weights, training code, and datasets to accelerate future research on agentic AI.
For more details, please refer to our [paper]() and [GitHub]().

# Model Downloads

| Model       |   Download   | Backbone Model                        | Licences|
| --------------------- | ------       | --------------------------- |--------------------------- |
| AFM-CodeAgent-7B-sft  | [🤗 **HuggingFace**]() |[Qwen-2.5-Coder-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct)  | Apache License 2.0|
| AFM-CodeAgent-RL-7B   | [🤗 **HuggingFace**]() |[Qwen-2.5-Coder-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct)  | Apache License 2.0|
| AFM-CodeAgent-32B-sft | [🤗 **HuggingFace**]() |[Qwen-2.5-Coder-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct) | Apache License 2.0|
| AFM-CodeAgent-RL-32B  | [🤗 **HuggingFace**]() |[Qwen-2.5-Coder-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct) | Apache License 2.0|
| AFM-MHQA-Agent-3B-sft | [🤗 **HuggingFace**]() |[Qwen-2.5-3B-Base](https://huggingface.co/Qwen/Qwen2.5-3B)  | Apache License 2.0|
| AFM-MHQA-Agent-3B-rl  | [🤗 **HuggingFace**]() |[Qwen-2.5-3B-Base](https://huggingface.co/Qwen/Qwen2.5-3B)  | Apache License 2.0|
| AFM-MHQA-Agent-7B-sft | [🤗 **HuggingFace**]() |[Qwen-2.5-7B-Base](https://huggingface.co/Qwen/Qwen2.5-7B)  | Apache License 2.0|
| AFM-MHQA-Agent-7B-rl  | [🤗 **HuggingFace**]() |[Qwen-2.5-7B-Base](https://huggingface.co/Qwen/Qwen2.5-7B) | Apache License 2.0|
| AFM-WebAgent-7B-sft   | [🤗 **HuggingFace**]() |[Qwen-2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) | Apache License 2.0|
| AFM-WebAgent-32B-sft  | [🤗 **HuggingFace**]() |[Qwen-2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) | Apache License 2.0|
| AFM-WebAgent-7B-rl    | [🤗 **HuggingFace**]() |[Qwen-2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) | Apache License 2.0|
| AFM-WebAgent-32B-rl   | [🤗 **HuggingFace**]() |[Qwen-2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) | Apache License 2.0|



# Data Downloads
TODO: add hf link after upload
- AFM-CodeAgent-SFT-Dataset
- AFM-CodeAgent-RL-Dataset
- AFM-WebAgent-SFT-Dataset
- AFM-WebAgent-RL-Dataset
- AFM-MHQA-SFT-Dataset
- AFM-MHQA-RL-Dataset

# License and Usage Information

## 1. Core License

This model is licensed under the **Apache License 2.0**, granting users the following rights:

✅ Commercial deployment

✅ Source code modification

✅ Patent authorization

✅ Closed-source derivatives

⚠️ Prohibition on using model names/logos for promotion without written authorization

⚠️ No warranties provided

## 2. Inheritance Declaration

This model is based on improvements from **Qwen2.5** (Apache 2.0 License). You must:

* Retain original Qwen copyright notices in derivative works.
* Clearly document changes made in modification notes.
* Adhere to any additional usage restrictions imposed by Qwen.