Model reference · open weights
Fara1.5 is an open-weight language model from microsoft, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
Fara1.5-9B [](https://aka.ms/fara1.5) [](https://github.com/microsoft/fara) [](https://huggingface.co/papers/2606.20785) [](https://aka.ms/fara1.5-9B-foundry) Fara1.5-9B is a multimodal computer use agent (CUA) for web browsers, from Microsoft Research AI Frontiers. It observes the browser through screenshots and acts on the user's behalf by emitting structured tool calls — click, type, scroll, visit URL, web search, and so on — to complete tasks end-to-end. The model is vision-only at perception time: it sees the browser through screenshots, not the DOM or accessibility tree. Internal reasoning and trajectory history are tracked as text. Given the latest screenshot and prior actions, it predicts the next action with grounded arguments (e.g., pixel coordinates for a click). Fara1.5-9B is supervised fine-tuned from Qwen3.5-9B on data generated by FaraGen1.5, our multi-agent pipeline that synthesizes web tasks, executes trajectories to solve them, and verifies the results before training. Please use only with the harness developped in Github/fara CLI or Magentic-Lite. It's co-designed with MagenticLite, and that's the recommended deployment for both research and production. Highlights - End-to-end web task completion. Fills forms, books reservations, applies for jobs, plans trips, runs shopping carts. Not just clicking around — sequencing actions toward a goal. - Vision-only perception. Operates on screenshots alone, no DOM access required. Matches the input modality available to a human user. - Coordinate-grounded actions. Predicts pixel-level click and drag targets directly. No separate grounding model needed. - Critical-points safety design. Trained to stop and ask before personal info entry, payments, submissions, sign-ins, sending messages, or other irreversible actions — even if it could technically continue. - 262K context. Long enough for multi-screenshot trajectories with full action history. - On-device viable. 9B parameters means it runs on a single A100/H100/B200 with room for the screenshot history. Model Details Recommended Deployment: MagenticLite The safest way to run Fara1.5-9B is inside MagenticLite, which provides: - Sandboxing — the browser run
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | microsoft |
|---|---|
| Type | Language models |
| Parameters (lead) | 9.4B |
| Variants | 3 |
| Runs with | transformers |
| Based on | Qwen/Qwen3.5-9B |
| Released | 2026-05-12 |
| Popularity | 6k downloads / month |
| Likes | 288 |
| Licence | Open weights |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
Using it via the API
Once AxForge deploys fara1-5 for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (fara1-5 below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/chat/completions \
-H "Authorization: Bearer $AXFORGE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"fara1-5","messages":[{"role":"user","content":"Hello"}]}'
Licence
Open weights under mit — commercial use is permitted. Deploy it on AxForge EU hardware on request. Read the licence ↗