1. Tutorial Introduction

Fara-7B is the first agentic SLM (Small Language Model) for computer use, released by Microsoft Research in November 2025.

Despite having only 7 billion (7B) parameters, the model performs exceptionally well in real-world webpage manipulation tasks, achieving top performance among models of similar size in multiple Web Agent benchmark tests.
It achieves state-of-the-art (SOTA) level performance and approaches or even surpasses larger-scale models on some tasks.

The relevant research results are Fara-7B: An Efficient Agentic Model for Computer Use.

Recommended hardware: Single RTX 5090 graphics card.

2. Project Examples

The following example demonstrates Fara-7B's automation capabilities on real web pages:

Prompt: Please look up Nobel laureate Tu Youyou for me.

Results page display: 

⚠️ All the above operations are performed through a real browser, not "text simulation".

---

3. Operation steps

1. After starting the container, click the API address to enter the Web interface

After starting the task in the OpenBayes console, click the API address in the right-hand information bar to access the Fara-7B Web UI page. 

If the page displays Bad Gateway  If you are unable to access the site, please wait 1–3 minutes and refresh the page.
This is a normal phenomenon during the initial loading and initialization of the model.

---

2️⃣ Using a Web UI for agent tasks

On the web page, you can directly input natural language tasks, for example:

• "Please search for tomorrow's weather in Beijing."
• Find the price of a product on different e-commerce platforms.
• "Find and organize information on the specified website"

The model will automatically:

1. Open webpage
2. Understanding page content
3. Perform click/scroll/input
4. Complete the task in multiple steps and provide the results. 

---

Output results page display: 

Project Support

Thanks to the Microsoft Research team for open-sourcing the Fara-7B project, and to the community contributors for their maintenance and improvements.

• Upstream warehouse:https://github.com/microsoft/fara
• Example of a major contributor: corbyrosset (commit history:https://github.com/microsoft/fara/commits?author=corbyrosset）

Citation Information

If you use this project in your research or paper, please cite the original work (as officially recognized):

@article{fara7b2025,
  title={Fara-7B: An Efficient Agentic Model for Computer Use},
  author={Awadallah, Ahmed and Lara, Yash and Magazine, Raghav and Mozannar, Hussein and Nambi, Akshay and Pandya, Yash and Rajeswaran, Aravind and Rosset, Corby and Taymanov, Alexey and Vineet, Vibhav and Whitehead, Spencer and Zhao, Andrew},
  journal={arXiv:2511.19663},
  year={2025}
}