charred wind
It seems like it is almost impossible to get any complex workflow done without hitting the rate limit. Finally I come down to a setup that it uses ChatGPT subscription as primary model provider using OAuth, another OpenAI API key as a fall back (unfortunately within the same account so it is same Org), and Kimi model as another fallback. It worked for a few days and now being rate limited again. The OpenAI message tells me to try in 6000 mins (what?) I looked in upgrading from Plus to Business plan but there are two problems: 1. The website wasn’t clear how much more quota I would be getting by upgrading. 2. The business plan requires a minimum 2 seats to be purchased so it is double the cost. Does anyone have a solid workaround (don’t tell me to run LLM locally). Thanks all!
⚠️ Agent failed before reply: All models failed (5): openai-codex/gpt-5.4: ⚠️ API rate limit reached. Please try again later. (rate_limit) | moonshot/kimi-k2.5: ⚠️ API rate limit reached. Please try again later. (rate_limit) | openai/gpt-5.1-codex: ⚠️ API rate limit reached. Please try again later. (rate_limit) | openai/gpt-5.4: ⚠️ API rate limit reached. Please try again later. (rate_limit) | openai/gpt-5.3-codex: ⚠️ API rate limit reached. Please try again later. (rate_limit).
Logs: openclaw logs --follow