Claude Opus 5.5 introduced
Claude Opus 5.5 is a more powerful, faster, cheaper, and safer model – an important announcement for AI users.
Today Anthropic released its new model Claude Opus 5.5. It came out about two months after Opus 5.
The new model is considerably more powerful. According to Anthropic, it delivers results close to Claude Fable 5.1, one of the strongest models on most tasks. At the same time, it is cheaper: $4 per million input tokens and $20 per million output tokens. In Opus 5 these prices were $5 and $25.
The model runs faster, uses fewer complex terms, and places the most important information at the beginning of the answer. It also performed well in tests: one tester transferred 680 k lines of code in less than a day. Usually this work would take a team several weeks.
From a safety perspective the model is also better than its predecessors and more resistant to being tricked with harmful instructions.
In the coming weeks Sonnet 5.5 and Haiku 5.5 are also expected to be released.
In short, Opus 5.5 is both stronger and cheaper. Good news for anyone working with AI.