Anthropic released Opus 5.5 on Tuesday, and the most immediate change is price: output tokens now cost $20 per million, down from $25. The model also runs faster and is aimed at the Claude users who rely on its top tier for coding and knowledge work.
Anthropic and Opus 5.5
Anthropic said Opus 5.5 set a new state-of-the-art in coding and knowledge work performance. It also said the model outpaces the larger Fable model in many benchmarks and succeeds in informal tasks that Fable failed to complete.
For users, that means the new tier is not just cheaper. It is also positioned as the strongest option in Claude for work that depends on code generation, technical retrieval, and structured outputs. Anthropic said Opus 5.5 is less likely to use jargon and more likely to put important information at the start of its messages.
Dario Amodei and prudence
The release lands two months after Opus 5 on July 24. It is Anthropic’s first model release since CEO Dario Amodei embraced calls to pace the frontier, even as the company pushed a more capable model into the market.
Amodei wrote earlier this month, “I have become convinced that fully addressing the risks requires even more prudence,” and also called for “not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.” That makes Opus 5.5 more than a routine refresh.
Anthropic said the model’s safety training was broadly similar to its predecessors. It also said Opus 5.5 had alignment testing and pre-release evaluation by outside organizations like METR and Frontier Design, while future models are already being prepared with more advanced training, evaluation, security, and monitoring systems.
Sonnet 5.5 and Haiku 5.5
Anthropic said Sonnet 5.5 and Haiku 5.5 will be released in the coming weeks. The company also said Opus 5.5 is comparable to Mythos in biology and cybersecurity capabilities, while the same safeguards limit how much the models can be used to discover exploits in compiled programs or develop recognizable biological weapons.
The unanswered question is how much of Opus 5.5’s benchmark edge over Fable will hold up in everyday use, where cost, speed, and output style matter more than lab scores. Users deciding whether to stay on Opus or wait for Sonnet 5.5 and Haiku 5.5 now have a cheaper top tier, but not a full picture of how that upgrade behaves outside Anthropic’s tests.
