Model guide / GPT

GPT guide

Connect a GPT-capable client through Stellahub’s OpenAI-compatible interface, then select the model ID returned for your API key.

OpenAI-compatible

Required access detailsStellahub API key · supplied base URL · model ID

Review environment variables →

01 Find the model ID

Model availability is determined by the authorization on your API key. Query GET /v1/models before sending a request:

GET /v1/models

curl "${STELLAHUB_API_BASE_URL}/models" \
  -H "Authorization: Bearer ${STELLAHUB_API_KEY}"

Copy an exact returned id into STELLAHUB_MODEL. If no GPT entry is returned, ask the organization that issued the key instead of guessing a model name.

02 Send a first request

Set STELLAHUB_MODEL to the returned ID, then choose a client example.

curl "${STELLAHUB_API_BASE_URL}/chat/completions" \
  -H "Authorization: Bearer ${STELLAHUB_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "${STELLAHUB_MODEL}",
  "messages": [{"role": "user", "content": "Give a one-sentence introduction to yourself."}],
  "stream": false
}'

Text chat uses POST /v1/chat/completions. Verify streaming, tool use, vision, reasoning, and context behavior against the returned model ID before relying on those features in production.

GPT availability notes

Use the Stellahub OpenAI-compatible gateway with the exact GPT model ID returned for your authenticated API key.

Run GET /v1/models first and copy an exact id from the response.
Tool use, streaming, vision, reasoning, and context limits depend on the returned model ID and your organization’s authorization.

Seeing 401, 403, 429, or an unavailable model? Review troubleshooting →