microsoft/FastContext-1.0-4B-SFT

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jun 14, 2026License:mitArchitecture:Transformer0.4K Open Weights Featherless Exclusive Warm

microsoft/FastContext-1.0-4B-SFT is a 4 billion parameter supervised fine-tuned repository-exploration subagent developed by Microsoft, built on the Qwen3-4B-Instruct backbone. This model is specifically designed to assist larger LLM coding agents by efficiently locating relevant code snippets within a repository. It excels at parallel read-only tool calls (READ, GLOB, GREP) to return compact file paths and line ranges, significantly reducing the main agent's token consumption and improving end-to-end resolution rates for coding tasks.

Loading preview...

FastContext-1.0-4B-SFT: Efficient Repository Explorer

FastContext-1.0-4B-SFT is a specialized 4 billion parameter subagent developed by Microsoft, designed to optimize repository exploration for larger LLM coding agents. It acts as a dedicated explorer, offloading the task of finding relevant code from the main agent, thereby saving tokens and improving overall efficiency. This model is built on the Qwen3-4B-Instruct backbone and has been refined through supervised fine-tuning (SFT) and reinforcement learning (RL).

Key Capabilities

  • Dedicated Repository Exploration: Functions as a subagent to perform read-only searches (READ, GLOB, GREP) within code repositories.
  • Parallel Tool Calling: Can issue multiple tool calls simultaneously to cover various search hypotheses.
  • Context Optimization: Returns compact file paths and line ranges as focused context, preventing the main agent's context from being polluted with irrelevant information.
  • Improved End-to-End Performance: Integrates with coding agents like Mini-SWE-Agent to boost resolution rates by up to 5.5% and reduce main-agent token consumption by up to 60%.
  • High Context Length: Supports a context length of up to 262K tokens.

Good For

  • Enhancing LLM Coding Agents: Ideal for developers looking to improve the efficiency and accuracy of their LLM-based coding assistants.
  • Reducing Token Usage: Significantly cuts down on the token budget spent by main agents on repository exploration.
  • Precise Code Citation: Generates accurate file and line range citations for relevant code snippets.
  • Complex Codebase Navigation: Excels at navigating and extracting information from large and complex code repositories.

Finetunes
1 models