Anthropic just changed the AI coding game.
Claude Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0, beating Opus 5.5 and Sonnet 5. It can even play Pokémon Red using screenshots alone. 30% faster. Up to 30% cheaper. The future of AI coding is here.
Claude Sonnet 5.5 is out.
30%+ faster than Sonnet 5, up to 30% cheaper for most work, and the same $2/$10 per million tokens. Anthropic's pick for everyday coding and docs. Opus 5.5 stays for the hard stuff.
World rank #1 Luna doing a 5-sense mental rehearsal simulation before hitting the Hard Enduro class at Red Bull Romaniacs.
Let's go...🏍️👷♀️ Failed 9 straight 720p 15-second clips. What is happening with Opus 5.5's performance?? I am exhausted...😅 AI Engine: Seedance 2.0 720p
Sonnet 5.5 and Opus 5.5 can write a trading bot in one prompt.
Neither can tell you if it makes money I typed one sentence into Horizon: "buy solana:HZ1JovNiVvGrGNiiYvEozEVgZ58xaU3RKwX8eACQBCt3 when RSI drops under 30, exit at +5%" it backtested the rules on years of history,
You have reached the end of the archive
All of Claude Opus 5