Model's context tier consumed, and AI-credit rate applied to OpenCode?

[undavide]

Hi,
I use Copilot (Enterprise account) as a model provider for OpenCode.

GitHub’s documentation says the 1M-token context option is available only in VS Code and Copilot CLI (for which I can choose default or long_context), but there's no such option in OpenCode.

I'm wondering whether Copilot serves the default context version for each model and whether there's a way to switch to the longer context explicitly. The cost varies, and I'm trying to be mindful of my spending.

Thanks!