Where do I see logs for my Model API requests?

Baseten doesn’t expose per-request logs containing Model API prompts or outputs. Model APIs follow a zero-data-retention posture for customer content, so Baseten does not persistently log or store request inputs or model outputs.

What you can see instead:

Aggregate usage via Model API usage reporting, including request counts and token usage over time, with grouping/filtering by dimensions such as API key, model, service tier, and user where available.

Operational metrics such as request counts/status rates through metrics export, which can help track 429, 5xx, and other response patterns.

Per-response usage details, where supported, in the API response itself, such as input/output token counts and cached-token counts.

If you need request-level application logs, log request IDs, timestamps, status codes, latency, and any prompt/output data you are permitted to retain on your own side of the call.