Anthropic introduced Sonnet 5.5 and made Claude Opus 5.5 the first model in its Claude 5.5 family. It says the new model costs 40% less to run than Opus 5, while output arrives more than 30% faster.
For teams paying for coding or agent-like work, that changes the math immediately. Anthropic is pairing lower runtime cost with higher five-hour usage limits on Pro, Max, Team, and seat-based Enterprise plans.
Claude Opus 5.5 pricing
Anthropic set Claude Opus 5.5 at $4 per million input and output tokens, with cache reads at $0.20 per million tokens. That pricing frames how quickly long chats, repeated retrieval, and tool-heavy workflows can add up for subscription users.
The company also said Claude Opus 5.5 performs at the level of Claude Fable 5.1 on most work. In practice, that means users are being asked to judge a lower-cost model against a stronger family name rather than against a simple price cut.
Frontier Design and METR testing
Before release, external evaluators including Frontier Design and METR tested the model. Anthropic also said it is the strongest-performing model it has tested to date on its automated behavioral audit, and that it is more resistant than Opus 5 to prompt injection.
That combination matters because the company is making a cost claim while also leaning on safety testing. The release is Anthropic's first since it called for pacing the frontier, and it says Claude Opus 5.5 comes with safeguards developed for its most capable models.
Life Sciences Verification Program
Anthropic said Claude Opus 5.5 is comparable to Claude Mythos 5.1 in biology and cybersecurity. Vetted organizations can apply now to the Life Sciences Verification Program to use it for biology research, and access to the Cyber Verification Program will expand in the coming weeks.
The practical question now is not whether the model is cheaper on paper. It is how much of that 40% drop users will actually feel once their own workloads, token counts, and five-hour limits are in play.
image_url
null
