Anthropic introduced Claude Opus 5.5, the first model in its new Claude 5.5 family, and said it costs 40% less to run than Opus 5 while matching Claude Fable 5.1 on most work. The release also brings higher five-hour usage limits for Pro, Max, Team and seat-based Enterprise plans, alongside a new rate limit reset that subscription users can save for later.
The launch matters now because Anthropic is trying to make its most capable model cheaper, faster and easier to use at the same time. At default settings, Claude Opus 5.5 will cost 40% less than Opus 5 on typical workloads, with input and output tokens priced at $4 and $20 per million, and cache reads priced at $0.20 per million tokens. Anthropic also said the model generates output more than 30% faster than Opus 5, a change that could matter as users push it through larger jobs rather than one-off prompts.
Before release, Claude Opus 5.5 was tested by external evaluators including Frontier Design and METR, and Anthropic said it is the strongest-performing model the company has tested to date on its automated behavioral audit. Early testers reported that it handled a 680,000-line code migration in less than a day, a task Anthropic said would have taken an engineering team weeks. In another test, Claude Opus 5.5 succeeded 39 of 40 times when asked to cut load times across every page of a web app, while Opus 5 made smaller improvements that also changed the app’s behavior.
Anthropic is pairing the cheaper pricing with broader access and tighter controls. The company said Claude Opus 5.5 includes the safeguards it developed for its most capable models, is much less likely than recent models to take hard-to-reverse actions or act outside its boundaries, and is more resistant than Opus 5 to prompt injection. It broadened alignment testing to cover longer tasks, impossible tasks and scenarios modeled on real incidents, which is the point at which performance claims start to run into model limits rather than erase them.
That split matters in the areas where the company is most cautious. Anthropic said Claude Opus 5.5 is comparable to Claude Mythos 5.1 in biology and cybersecurity, and that it is deploying the model with safeguards similar to those on Claude Fable 5.1. Vetted organizations can apply to Anthropic’s Life Sciences Verification Program for biology research use, while verified cybersecurity practitioners will be able to use the model for their work when access expands through the Cyber Verification Program in the coming weeks. The result is a launch that widens the circle of users even as Anthropic keeps the most sensitive uses behind verification and makes clear the model is stronger, but not unrestricted.
For subscription users, the practical change is immediate: more time with the model inside each five-hour window, plus a reset they can hold for later. For Anthropic, the bet is that lower cost, faster output and broader access will make Claude Opus 5.5 the default choice for heavier work, even as the company continues to draw lines around the tasks it thinks need extra scrutiny.
