Opus 4.7

1. I built a free continuous Claude benchmark - Opus 5 currently ranks #1 across 22 active models (www.reddit.com via reddit)

   3w opus

   https://preview.redd.it/xcejpnf5vamh1.png?width=1903&format=png&auto=webp&s=a53a982f197cb7128f5b862cda4e8e1e8a3aaeb6 I built AIStupidLevel, a free-to-try platform that continuously benchmarks Claude and other leading LLMs across coding, de…

2. It can't all be downhill. (www.reddit.com via reddit)

   3w opus

   Every time there's a new Opus model, everyone complains that it's worse than the previous one. And yet, I notice that people usually suggest the alternative of using the previous Opus model.

← all threads