I think this image explains it best: https://www-cdn.anthropic.com/images/4zrzovbb/website/1f044104447e9db6b22db3a06e45d114f50f274e-2200x1174.png https://www-cd

I think this image explains it best: https://www-cdn.anthropic.com/images/4zrzovbb/website/1f044104447e9db6b22db3a06e45d114f50f274e-2200x1174.png https://www-cdn.anthropic.com/images/4zrzovbb/website/1f0441... I.e. Opus is the largest and best model of each family but Sonnet is the first model of the 3.5 family and can beat 3's Opus in most tasks. When 3.5 Opus is released it will again outpace the 3.5 Sonnet model of the same family universally (in terms of capability) but until then it's a comparison of two different families without a universal guarantee, just a strong lean towards the newer model.