- Endpoint:
POST /v1/responses - Model IDs:
gpt-5-6-luna,gpt-5-6-terra,gpt-5-6-sol - Authentication:
Authorization: Bearer <POYO_API_KEY>
| Field | Requirement | Constraint |
|---|---|---|
| model | required | gpt-5-6-luna, gpt-5-6-terra, or gpt-5-6-sol. |
| input | required | String or Responses-style input items. |
| instructions | optional | High-level system instructions. |
| max_output_tokens | optional | Maximum output tokens. |
| reasoning | optional | Reasoning configuration for supported variants. |
| stream | optional | Enable streaming. |
- Keep API keys in a server-side secret manager.
- Validate user input before calling the API.
- Apply request timeouts and retry only rate limits or transient 5xx responses.
- Bound output tokens and log usage for cost control.
- Redact sensitive prompt and response content from logs.
- Recheck the live model page before hard-coding prices.
Treat non-2xx HTTP responses and non-success API code values as errors. Preserve the HTTP status, API code, request ID when present, and a redacted response body for debugging.
Schema: PoYo API documentation