- Check
GET /v1/modelswith the samegd-…key. - Copy one of the four Kimi model IDs without changes.
- Make a minimal non-stream request to
POST /v1/chat/completionswith a smallmax_tokens. - After a regular response, test
stream: true. - For 429, read the exact
error.code: causes and retry timing differ.
| Symptom | Check |
|---|---|
404 model_not_found | exact model ID with no aliases or spaces |
402 request_reserve_exceeds_credit | the reserve also covers the request input; check the request size and max_tokens — without it the output part uses 32,000 |
| Usage appears twice in a stream | this is native choice usage plus the separate frame with include_usage; do not sum them |
429 rate_limit_exceeded | wait 60–90 seconds; this is not exhausted plan quota |
429 kimi_queue_full | retry with a delay |
503 kimi_unavailable | retry later |
404 file_not_found for a fresh file | a file lives 24 hours and is visible only to the key that uploaded it |
415 unsupported_file_type on upload | eight formats are supported; images, audio, and video are not accepted |
400 invalid_managed_service_request | the search and fetch body must carry exactly one field — query or url — and the body itself must be valid JSON |
422 managed_service_request_failed | the site itself, or the search itself, refused. Repeating is pointless and there is no Retry-After: use another address |
402 insufficient_credit on DELETE /v1/files/{file_id} | files are free, but zero remaining credit closes deletion too; top the key up, or wait for expiry after 24 hours |
A second input_file with the same file_id vanished | the substitution happens once per request; to point at the same document again, say so in words in your own text |
| The 50 MiB quota ran out on a couple of files | space is counted on the encrypted record together with the extracted text, which for plain text is about 3.6 times the file size |
Send the [code · req_…] tail from the error text (or the request_id), the endpoint, model id, stream true/false and the exact time with its zone. Do not send the full key, the prompt or the answer.