even grok 4.1-fast is bad i tested

[keen swallow]

i read so many different opinions on this. can you elaborate what was bad please?

@oblique inlet

[oblique inlet]

@keen swallow Example: when i tried Ith Claude Sonnet 4.5, he was much more autonomous and could fix webcam connexion easily and Slack Connexion and so many things while grok4.1 had to prompted again and again and couldn’t resolve some problems on its own

[keen swallow]

That sounds bad indeed. Gives me the feeling I had with gpt 5 mini, same frustration at times. The only thing that i am a little afraid of are the costs. OpenClaw is insane in token usage.. how has it been for you costwise on sonnet 4.5?

[oblique inlet]

@keen swallow it did cost 30 dollars for like 4 hours of continuous use