Text models
Connect and verify Gemini 4 Carbon
Check the model ID, text requests, streaming, and recorded usage.
Verify the model route
Look up gemini-4-carbon in GET /v1/models and check its availability before sending a request.
Start with a short request
Send contents in the request body and verify the response shape. Use streamGenerateContent for SSE and record final usage after the stream closes.
Run a reproducible regression
Keep a small group of short tasks and another group of longer tasks. Record output completeness, tool calls, actual tokens, and billing IDs.
Example request
curl "https://api.apimodel.store/v1beta/models/gemini-4-carbon:generateContent" \
-H "Authorization: Bearer $MODELPORT_API_KEY" \
-H "Content-Type: application/json" \
--data '{"contents":[{"role":"user","parts":[{"text":"List three acceptance checks for an offline notes app."}]}],"generationConfig":{"maxOutputTokens":256}}'