Github Copliot provider allows only 128k token conversations.

I have noticed that when using github copilot as my provider i get this error for prompt longer that 128k tokens. this seems weird, since i am using gemini 2.5 pro as my model, but conversations cannot be more than a few messages long because token limit is reached. is there a way to fix that?

This issue might be a duplicate of existing issues. Please check:

応答なし microsoft/vscode-copilot-release#1551: Similar token limit error with different models - shows how OpenCode handles token limit errors when context exceeds provider limits
microsoft/vscode-copilot#6117: Copilot models respond with Bad Request errors after chat - similar pattern of GitHub Copilot provider issues after conversation continues
Chat Not Working With Code-Server (1.80) microsoft/vscode-copilot-release#353: Copilot Gemini 2.5-pro authentication/compatibility issues - shows historical issues with Gemini models via GitHub Copilot provider

Feel free to ignore if none of these address your specific case.

I think we need to update models.dev to reflect this

Made issue on models.dev to track: anomalyco/models.dev#172

I have encountered the same problem.

I don't see any options for configuring manually context limits, which surprised me. Does this exist?

@astrolox

@astrolox there are options for this actually:

"provider": {
    "github-copilot": {
      "models": {
        "gemini-2.5-pro": {
          "limit": {
            "context": 128000,
            "output": 65536
          }
        }
      }
    }
  }

We will need to fix models.dev to accurately portray the limits but you can do this^ as a workaround, annoyingly github copilot doesn't exactly outline the limits for each model

Thanks for the help, and for the fast response I will implement this immediately! 💪

Makes GitHub Copilot models effectively unusable 😿

@amal

@amal I made an issue on models.dev to update the models. if it doesn't have any open prs I will try to jump on that next sorry about delay on this I mentioned to team a few weeks ago

I made a PR to models.dev and pinged team to review hopefully it gets shipped soon

Try restarting opencode or running:

rm ~/.cache/opencode/models.json

and it should be fixed

Navigation Menu

Repository navigation

Github Copliot provider allows only 128k token conversations. #2317

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Github Copliot provider allows only 128k token conversations. #2317

Description

Activity

github-actions commented on Aug 29, 2025

rekram1-node commented on Aug 29, 2025

rekram1-node commented on Aug 29, 2025

astrolox commented on Aug 30, 2025

rekram1-node commented on Aug 30, 2025

Steinvorth commented on Sep 1, 2025

art-shen commented on Sep 10, 2025

rekram1-node commented on Sep 10, 2025

rekram1-node commented on Sep 11, 2025

rekram1-node commented on Sep 11, 2025

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions