First check which models your GitHub account can use.
Open GitHub Copilot Features settings ↗ to view and manage model availability. Business or Enterprise organization policy may further restrict it.
1. List and inspect models
Shell
curl http://127.0.0.1:8313/v1/models
curl http://127.0.0.1:8313/v1/models/full/The first response is a simplified OpenAI-style list. The second preserves full Copilot model metadata for inspection and configuration.
2. Important fields
| Field | Use |
|---|---|
id | The model ID sent by clients. |
supported_endpoints | Select Responses, Chat Completions, or Anthropic Messages. |
capabilities | Inspect context, output, vision, tools, streaming, and other reported capabilities. |
capabilities.supports.reasoning_effort | Only send effort values present in this array. |
The full schema is not a stable public standard.
Scripts should tolerate missing fields and use conservative defaults.
3. Filter with jq
Shell
curl -s http://127.0.0.1:8313/v1/models/full/ |
jq -r '.data[] | [.id, (.supported_endpoints // [] | join(","))] | @tsv'
curl -s http://127.0.0.1:8313/v1/models/full/ |
jq -r '.data[] | select((.supported_endpoints // []) | index("/responses")) | .id'4. Match clients to protocols
- Claude Code: select a model with
/v1/messages. - Codex: select a model with
/responses. - pi and OpenClaw: configure the adapter that matches the model endpoint.
The local API response best represents what is callable for this account and organization right now.
GitHub Copilot supported models ↗
Next: configure your client
After confirming the model and endpoint, return to the guide index and choose Claude Code, Codex, pi, or OpenClaw.
View all guides →