Claude Opus 4.7 Intelligence, Performance and Price Analysis

I am planning to cancel my Claude sub after I've used up remaining tokens this weekend. Model is fine but the harness and reliability have absolutely nosedived. Just unacceptable for the high price.

Vibe-coded way too much of their own infra.

Interesting site - going to delve. Especially 4.7 stuff....

been waiting for the artificialanalysis numbers on this one. the speed drop vs 4.5 is more than i expected tbh. anyone seen a side-by-side on coding tasks specifically?