@repligate @MoonL88537 In those tests I was trying out non-thinking models, so that was non-thinking Sonnet 3.7 w/ a CoT prompt (it loops more than other Sonnets).

But on your suggestion I just tried Sonnet 3.7 thinking and it's completely unhinged! >5k or 6k tokens of output
https://t.co/mFEelWUXKL