#(all profiles unavailable) (format)
1 messages · Page 1 of 1 (latest)
(format) is OpenClaw’s error category, not message formatting.
In that line:
Provider openai is in cooldown (all profiles unavailable) (format)
it means OpenClaw saw recent request-shape/validation errors for that provider (typically HTTP 400/422, e.g. invalid field/tool-call payload), so it marked OpenAI profiles as temporarily unavailable and skipped them.
Your second error is separate:
Unknown model: claude-cli/opus-4.6 (model_not_found)= that model id/provider combo isn’t recognized in current config/runtime.
If you want, I can help you debug both quickly — best next data is:
openclaw --version- one log line around the first 400/422 error (it usually names the bad field)
- Latest Version
- 400 Unsupported value: 'none' is not supported with the 'gpt-5-mini' model. Supported values are: 'minimal', 'low', 'medium', and 'high'.
Yep — that log confirms it.
(format) here means “request payload format/validation error” (OpenClaw failover reason), and in your specific case it’s:
Unsupported value: 'none' ... Supported values: minimal|low|medium|high
So OpenAI rejected the request because thinking was set to none for that model, and OpenClaw put that provider into temporary cooldown.
Quick fix:
- Send:
/think minimal(or/think low) - Check current setting:
/think - Retry your message
Also in your original error, claude-cli/opus-4.6 is a separate issue (model_not_found) — use /model list and pick from what your runtime actually has.
Docs: https://docs.openclaw.ai/tools/thinking and https://docs.openclaw.ai/concepts/models
Glad that helped! 🎉 If this fixed it, please right-click the solution message → Apps → ✅ Mark Solution.
No available auth profile for openai (all in cooldown or unavailable)
Yep — that means you’re now in the cooldown phase from those earlier 400s.
Since your OpenAI request used unsupported thinking (none), OpenClaw marked the OpenAI profile(s) temporarily unusable, so now you get:
No available auth profile for openai (all in cooldown or unavailable)
Do this:
- Set thinking to a supported level:
/think minimal(or low/medium/high)
- Retry after cooldown expires (it auto-recovers).
- If you want to verify cooldown state/expiry:
openclaw models status --json- check
auth.unusableProfilesfor OpenAI entries.
Also keep in mind your fallback model in that error (claude-cli/opus-4.6) is invalid (model_not_found), so use /model list to pick a valid configured model.
can I set think level in config files?
Yes 👍
Set it in config with thinkingDefault:
{
agents: {
defaults: {
thinkingDefault: "minimal"
}
}
}
You can also override per agent via agents.list[].thinkingDefault.
Valid levels: off | minimal | low | medium | high | xhigh | adaptive