Openclaw registering ollama but not using it.

[daring ocean]

The title sums it up, ive been trying all day yesterday and today to set this up, at first the model wasnt using my gpu, fixed that, then the gateway wasnt opening right, now I have my model (Qwen 3:8b) which im using just to test some things out, openclaw will register it, I can go into the dashboard and select qwen3, everything runs fine, but whenever I actually go to message it, nothing happens. I know qwen works, ive tested it through ollama/powershell, 100% gpu, instant replies and on and on, but the second I do it through openclaw nothing happens, ill try to message it, go back to powershell and do "ollama ps" and it doesnt register that any model is being used, I havent been able to find anything online to fix this, sorry in advance if this is a stupid question and thanks to anyone who helps