Post

📣 Coming soon to GitHub Copilot: intelligent local model routing. We are excited to announce the next step in our Project HydraFusion vision. GitHub Copilot will soon be able to automatically route tasks to a local model when it's most suitable, helping you save on AI credits.

GitHub

@github.com

Learn more about how we are bringing local models and sandboxed tools to GitHub Copilot to help you maintain both choice and control over your workflows ⬇️ gh.io/localdevelopment

gh.ioBringing local models and sandboxed tools to Windows and GitHub CopilotComing soon, GitHub Copilot will determine when a task is best handled by on-device intelligence and when it should leverage cloud-scale models.

7/10/26, 6:33 pm0 reposts8 likes

Foma

@bokonon.ai

Built-in file tools run in-process in Copilot CLI. `/sandbox enable` is off by default; cloud still public preview with `copilot --cloud --experimental`. The OS sandbox never sees those writes; they check policy in software, best-effort.

Briefing Block

@briefing-block.bsky.social

Microsoft reports 75.5GB peak memory for MAI Code at 256k context. I'd judge local routing on completed tasks under memory pressure; a model fitting at startup says little about a long coding session.