Looks marginally better than Opus 5.5 on targetted tasks. Let's wait for Artificial Analysis to confirm how good it is for most users and if it outperforms Opus 5.5 on Terminal V4.0 and Fronttier Code 2.0.
Its still much more expensive than Opus 5.5, for a heavy user, its still unclear what value it serves.
It is far, far better on reasoning intensive tasks. If you don't need it, great, but if you do, it's a godsend. It saves me from typing thousands of lines of prompting to opus 5.5.
After seeing the benchmarks on AA and running some tests, I've to agree to with, its definitely good, I'd still preserve it for most complex tasks if I have limited budget for AI tools. Otherwise its more than capable of being a good all rounder as well.
3 comments
[ 0.30 ms ] story [ 17.0 ms ] threadIts still much more expensive than Opus 5.5, for a heavy user, its still unclear what value it serves.