GitHub Copilot Adds Intelligent Local Model Routing (Project HydraFusion)
Melvin Vivas · X video post · 2026-10-08 · 2:33 · 309 views · Open on X
Topics: AI Dev Tools & Productivity, LLMOps, Deployment & Monitoring, Industry Trends & Job Market · Level: beginner
Summary
This announcement video explains that GitHub Copilot will soon send some tasks to local models on capable machines, building on Project HydraFusion, its system for routing work across many cloud models. Copilot will find models you installed with Ollama or Microsoft Foundry Local, and it will offer MAI Code 1.1 Flash as a local model. In auto-routing mode it uses these local models for simple, long-running or offline work, so those tasks use no AI credits.
Key points
- Project HydraFusion routes planning, implementation and execution across many cloud models. Microsoft says this gives frontier-level performance and cuts cost by up to 67%.
- Copilot will automatically find local models installed with supported providers (Microsoft Foundry Local, Ollama), and they will appear in the model picker. You won't need to set them up inside Copilot.
- On capable devices you can install MAI Code 1.1 Flash, a local coding model, directly from GitHub Copilot.
- In auto-routing mode, Copilot sends simple tasks to a local model, such as explaining how a codebase works or how to run a project.
- A suggested pattern is to plan with a large frontier model (e.g. Astra), then hand the implementation to a local model that can run overnight.
- Copilot automations (scheduled background tasks) suit local models well: they can run very long tasks and use zero AI credits.
- Local models keep working offline, so you can still code with AI when you have no internet connection.
- Project HydraFusion is available now under Experimental Settings in GitHub Copilot. Local routing is coming soon.
Resources mentioned
- GitHub Copilot · tool · github.com · free · recommended by both Bashiri Smith & Melvin Vivas
One of the agentic coding tools the guide names as letting non-technical builders ship.
Also in: Microsoft MAI-Code-1.1-Flash: On-Device Coding Model in GitHub Copilot (Melvin Vivas on X · notes), Opinion: Enterprises Pick GitHub Copilot, but Cursor Is Better (Melvin Vivas on X · notes) - Project HydraFusion · tool · github.blog · paid
Copilot's model orchestration feature that plans, implements and executes tasks across many cloud models to cut cost. - Microsoft Foundry Local · tool · github.com · free
Microsoft's tool for running AI models locally on your own device. - Ollama · tool · x.com · free · recommended by both Bashiri Smith & Melvin Vivas
Open-source tool for downloading and running LLMs locally from a simple command line or API.
Also in: Run Local LLMs on the Computer You Already Have: A Beginner's Path (Melvin Vivas on X · notes), Running LLMs Locally Without an Expensive Rig (Melvin Vivas on X · notes), Adding Vercel AI Gateway as a provider in AIBackends with Devin (Melvin Vivas on X · notes), Running Qwen3.8-27B Locally on an M5 Max MacBook with Inco Splash (Melvin Vivas on X · notes) and 14 more - MAI-Code-1.1-Flash · tool · microsoft.ai · free
Microsoft's local coding model, which you can install directly from GitHub Copilot on capable devices.
Also in: Microsoft MAI-Code-1.1-Flash: On-Device Coding Model in GitHub Copilot (Melvin Vivas on X · notes) - GPT-6 Astra · tool · openai.com · paid
The model announced in the quoted launch post, pitched as the developer's most capable model for work, coding, science and cybersecurity, and able to operate a computer.
Also in: Glimpse: A Visual Reasoning Model That Uses a Visual Scratch Space (Melvin Vivas on X · notes), Is AI Getting Cheaper? How to Measure LLM Cost Per Task and Per Intelligence Level (Melvin Vivas on X · notes), Comparing Opus 5.5, GPT-6 Astra and Grok 4.7 on OpenRouter (Melvin Vivas on X · notes), OpenAI Makes GPT-6 Astra and GPT-6.1 Sol About 50% Faster (Melvin Vivas on X · notes) and 55 more
Try this
- Turn on Project HydraFusion in GitHub Copilot under Experimental Settings.
- Install local models with Ollama or Microsoft Foundry Local so Copilot can find them automatically.
- Set Copilot to auto-routing so it can send simple tasks to local models.
- Plan with a frontier model, then hand the implementation to a local model to save credits.
- Use local models for Copilot automations (scheduled background tasks) and for offline work.
More in AI Dev Tools & Productivity
- Letting coding agents set up and run GitHub repos for you
- Why Model Speed Matters in Coding Agents: Codex vs ChatGPT Plans
- Claude in Google Docs, Sheets and Slides: An Alternative to Gemini
- Local $0 Video Head Swap: MiniMax H3 Ref2VA (INT8) + LoRA in ComfyUI
- Pairing Grok Bot with Cursor Cloud Agents for productivity
- Running Grok Bot with ComfyUI locally from an iPhone