GitHub launches Copilot Fusion to route and combine coding models
Original title: GitHub Launches Copilot Fusion to Automatically Select and Combine AI Models for Coding
Fusion is generally available on all Copilot plans, bills each model at its normal rate, and offers a 30% discount through November 30, 2026.
GitHub released Copilot Fusion, a runtime router that picks models and execution patterns for coding, debugging, and review instead of requiring a manual choice per step. It is generally available in Visual Studio Code, Copilot CLI, the GitHub Copilot app, and JetBrains IDEs, subject to subscription and organization policy. Standard, High, and Max modes trade latency for deeper review; a Fast mode is limited to VS Code inline chat. Discarding a draft does not undo file edits that draft already made.
Unconfirmed: GitHub's cited prompt-injection and Agentjacking reports are described as separate issues, not flaws shown in Fusion.
- Single, Cascade, and Critique patterns can change between prompts in one conversation.
- Local routing preview can use MAI Code 1.1 Flash or Qwen3.8 Flash Next.
- Fusion disappears from the picker if policy leaves no eligible models.
- There is no separate Fusion fee; High and Max can use more credits.
- Fusion does not launch subagents and has no independent context window.
- Vendors
- GitHub
- AI models
- MAI Code 1.1 FlashQwen3.8 Flash Next
From the original article
2 min read at gbhackers.com
GitHub has launched Copilot Fusion, a runtime model router and orchestrator that automatically selects AI models and execution patterns for coding tasks. Instead of requiring developers to choose a model for each step, Fusion coordinates reasoning, code generation, debugging, and review, ultimately providing a single cohesive response. The feature is available across all Copilot plans.
By Divya · gbhackers.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.