All new model are timeout

New model like glm 5.1, kimi 2.6,… are always time out

So I hope we can fix it

glm works just fine, i have been having issues with recent deepseek release and kimi-k2.6, please review and fix

it’s because people running Openclaw are pulling in 1000’s of requests per minute

But I have problem with it