Add support for GitHub Copilot models 1 million token context window

[GianKyle]

What do you want to change?

By default, github copilot models are set to previous context window limit (For gpt-5.5, it was at 400k, now it's at 1m). I want to be able to select not only the model's thinking mode (high, xhigh, etc.) but also context window (400k, 1m, depending on model).

Why?

GitHub just announced this recently: https://github.blog/changelog/2026-06-04-larger-context-windows-and-configurable-reasoning-levels-for-github-copilot/

How? (optional)

No response

[github-actions]

This issue was auto-closed. All issues from new contributors are auto-closed by default.

Maintainers review auto-closed issues daily and reopen worthwhile ones. Issues that do not meet the quality bar in CONTRIBUTING.md will not be reopened or receive a reply.

If a maintainer replies lgtmi on one of your issues, your future issues will stay open. If a maintainer replies lgtm, your future issues and PRs will stay open.

See CONTRIBUTING.md.

[GianKyle]

Hi — can this issue be looked on? :)

[milespossing]

@GianKyle it's far from perfect (suggestions welcome!) but I fixed this with a simple extension which I had an agent write for me:

milespossing/pi-copilot-discovery

---

There are still a few issues. I'm not sure if it's me or pi, but tokens don't seem to last very well, even in running sessions. It also doesn't seem to be possible to get the different context windows from the copilot service (in copilot cli I seem to be able to access different context window sizes, that isn't reflected in the public api). Until it's perfect I wouldn't want to upstream it, but it might be worth doing at some point