The One API now shows model status and pauses slow models
The One developer API now lists dozens of named models, shows whether each is working, and checks their speed all day. A model that runs slowly is paused and comes back on its own once it is fast again.
What changed
- The model list covers Claude, GPT, Gemini, Grok, Qwen, Kimi, GLM and DeepSeek models, with One Auto first.
- Every model shows a status: working, slower than usual, or down.
- A speed check runs every 30 minutes. A model that is slow twice in a row is paused, and it returns after two fast checks.
- Asking for a paused model by name gives a clear message that points to One Auto or another model.
- Error messages are now plain, and bad settings such as an impossible temperature are rejected with a reason.
Why
If you build on an API, a slow model is worse than a missing one. One Auto now skips paused models for you, so your app keeps answering quickly.