Model status

Check Claude, GPT, Gemini and other text models on API Route: sampled availability, time to first text, first-attempt success rates and probe history.

About these checks

API Route sends short text requests through its public API. The table covers text models in families such as Claude, GPT, Gemini, DeepSeek and Qwen; the live list and check interval above reflect the current configuration. Image, audio and video generation are not tested.

Green means a recent check passed without a degradation condition. Yellow includes failed first attempts, repeated slow responses and initial recovery. Red means two consecutive rounds failed both the first attempt and retry. Missing, expired or monitoring-error results are unknown. Expand a model to view its history; individual checks are retained for 90 days.

Common questions

Is this the model provider’s official service status?

No. These are samples of the API Route request path from our monitoring location. They do not represent a provider’s entire service, every upstream channel or every user region.

Why can my request fail when a model is green?

A successful short text check is not a guarantee for every request. Timing, routing, account limits, prompt length and request parameters can differ. Check your request log and the model’s recent history when troubleshooting.

What do first-attempt success, retries and response times mean?

First-attempt success is the share of valid rounds whose initial probe request succeeded. Missing samples and monitoring errors are excluded; a successful follow-up retry does not turn that first attempt into a success. Each probe request may include gateway routing retries. Time to first text ends at the first visible text; total duration ends when the request completes or fails. These are short-probe timings, not a reasoning benchmark.

Related pages

Model Square

Browse available API-Route models, supported groups, and current API rates.

Blog