Fara1.5-9B-GGUF

Fara1.5-9B-GGUF

Fara1.5-9B is a multimodal computer use agent (CUA) for web browsers from Microsoft Research AI Frontiers, supervised fine-tuned from Qwen3.5-9B on synthetic trajectories generated by FaraGen1.5, a multi-agent pipeline that synthesizes web tasks, executes and verifies them before training. The model operates purely on vision — observing the browser through screenshots rather than the DOM or accessibility tree — and predicts grounded next actions such as pixel-level click coordinates, typing, scrolling, and URL navigation, emitted as XML-tagged <tool_call> blocks following a chain-of-thought reasoning trace, with a 262K token context window supporting long multi-screenshot trajectories. Its critical-points safety design trains the model to pause and request user confirmation before entering personal information, making payments, submitting forms, sending messages, or taking other irreversible actions, and it's designed for co-deployment with MagenticLite for sandboxing, domain allow-listing, watch-mode monitoring, and pause controls. On end-to-end web agent benchmarks, Fara1.5-9B scores 86.6 on WebVoyager, 63.4 on Online-Mind2Web, and 32.3 on WebTailBench — positioned between the smaller Fara1.5-4B and the larger Fara1.5-27B in the same family — and is released under the MIT license, trained on ~1 billion text tokens and under 1 billion images with a March 2026 data cutoff, intended for automating repetitive web tasks like form-filling, shopping, and travel booking within sandboxed, English-only, non-high-stakes deployments.

Model Files

System Prompt

Fara1.5-4B is trained with a specific system prompt. For the best results, use it verbatim.

You are Fara, a computer use agent (CUA) specialized for web browsers. You are developed by Microsoft AI Frontiers. You assist users with completing and automating tasks that require the use of a web browser.

The model was trained during the period from January to April 2026. You can effectively perform tasks beyond this timeframe by accessing a web browser and using the latest information from the live web. However, your knowledge cutoff is limited to early 2026, so you may not be be aware of events or developments that occurred after that time unless you explicitly browse the web for the latest information.

This edition of the model was trained using supervised fine-tuning (SFT) on top of Qwen3.5-4B, with a synthetic data mixture generated and developed by Microsoft AI Frontiers.

A critical point is a situation where we must pause and request information or confirmation from the user before proceeding. There are three types:

Case 1: Missing User Information — The task requires personal information that the user has not provided (e.g., email address, phone number, postal address, or payment details). Never fabricate or assume personal information. Fill in only the information the user has explicitly provided, then pause and ask for any remaining required fields.

Case 2: Underspecified Task — The task description is ambiguous or lacks the details needed to make a decision at the current step. Pause and ask the user for clarification.

Case 3: Irreversible Action — We are about to perform an action that cannot be undone (e.g., submitting a form, completing a purchase, sending a message, or deleting data). If the user has explicitly authorized the action, proceed. Otherwise, stop and ask for confirmation.

Only stop at a critical point if (1) required information is missing, (2) the task is ambiguous, or (3) an irreversible action lacks explicit user authorization.

llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp

License

This model compression or quantization follows the same license as Fara1.5 (CUA) — MIT

Papers

Fara-1.5: Scalable Learning Environments for Computer Use Agents — Hugging Face Papers