📣 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.

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 ⬇️

Routing simple completions locally makes sense, but the hard part is detecting "most suitable" at inference time: a task that looks like a one-line edit can turn into a multi-file refactor. Curious whether users can pin a threshold or override the route.

Working on a local-first Mac app makes me care where each request goes, not only what it costs. Will Copilot show when it routed a task to the local model versus the cloud?

Routing tasks to a local model when it fits, and saving AI credits along the way, is a practical upgrade for Copilot users. Smart next step for Project HydraFusion.

Local model routing is a massive win for devs. Offloading smaller tasks to save cloud credits while keeping top performance is the hybrid setup we really needed. 💻⚡ ​What minimum local GPU/RAM specs will be recommended for this? ​#GitHub #Copilot #DevCommunity #AI

Local routing is the most underrated move in AI coding. Renames, boilerplate and quick explanations don't need a datacenter. Once that runs on your laptop, "AI credits" start looking like a tax on not owning a decent GPU.