GitHub Copilot is also counting the supposed free gpt-5-mini in the premium-request usage.

[jackywu]

After I installed copilot-api v0.7.0 (http://192.168.50.1:4141), I called it with opencode using gpt-5-mini. After using it for some time, I checked the copilot usage in vs code and found that the premium request usage was continuously increasing. According to the description on https://github.com/features/copilot/plans, the pro plan has unlimited usage rights for gpt-5-mini, and I am on the pro plan. How can I find the cause of this issue?

By the way, I called MCP through LLM, not sure if this behavior is related to the problem.

    "copilot-proxy": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Copilot Proxy",
      "options": {
        "baseURL": "http://192.168.50.1:4141"
      },
      "models": {
        "gpt-4.1": {
          "name": "gpt-4.1"
        },
        "gpt-4o": {
          "name": "gpt-4o-2024-11-20"
        },
        "gpt-5": {
          "name": "gpt-5"
        },
        "gpt-5-mini": {
          "name": "gpt-5-mini"
        },
        "gpt-5.1": {
          "name": "gpt-5.1"
        },
        "gpt-5.2": {
          "name": "gpt-5.2"
        },
        "gemini-3-flash": {
          "name": "gemini-3-flash-preview"
        },
        "gemini-3-pro": {
          "name": "gemini-3-pro-preview"
        },
        "raptor-mini": {
          "name": "oswe-vscode-prime"
        },
        "claude-haiku-4.5": {
          "name": "claude-haiku-4.5"
        },
        "claude-sonnet-4.5": {
          "name": "claude-sonnet-4.5"
        },
        "claude-opus-4.5": {
          "name": "claude-opus-4.5"
        }
      }
    }

[caozhiyuan]

@jackywu Are you using opencode? Refer to this configuration #167 (comment) . "npm": "@ai-sdk/openai-compatible" does not support model reasoning functionality; you need to use the message-api, which supports reasoning for gpt/claude/gemini.