Which API should you use with GitHub Copilot?

For long GitHub Copilot agent sessions in VS Code, add Standard Compute as a Custom Endpoint model (base URL https://api.stdcmpt.com/v1) and those sessions run on a flat plan from $19 a month instead of Copilot's own models. Copilot meters chat, agent mode, code review and Copilot CLI against monthly GitHub AI Credits, so long agent runs are what drain a plan first. VS Code's docs say bring-your-own-key models work even without a Copilot plan; inline code completions are the part that still needs a GitHub account.

Light or occasional use

If your agent sessions fit inside Pro's monthly credits, the $10 plan is hard to beat, and it covers inline completions as well. Watch the credit meter through a busy week before buying extra credits.

One fixed monthly bill

Keep Copilot Free or Pro for completions and pick the Standard Compute model in the chat picker for long agent runs. Agents only list models with tool calling enabled, which the config below turns on.

API and subscription options for Copilot

Prices and allowances can change. Copilot prices and credit amounts were read from github.com/features/copilot/plans on October 13, 2026. The Custom Endpoint steps, the no-plan BYOK statement and the completions caveat come from the VS Code language models documentation. Copilot plans · VS Code: bring your own model. Spot an error? Tell us.

Add a Custom Endpoint model in VS Code

[
  {
    "name": "Standard Compute",
    "vendor": "customendpoint",
    "apiKey": "${input:standardComputeKey}",
    "apiType": "chat-completions",
    "models": [
      {
        "id": "standardcompute",
        "name": "Standard Compute",
        "url": "https://api.stdcmpt.com/v1/chat/completions",
        "toolCalling": true,
        "vision": false,
        "maxInputTokens": 200000,
        "maxOutputTokens": 32000
      }
    ]
  }
]

Restart VS Code if the model does not appear in the picker. The input variable keeps the key out of the file.