Anthropic releases Claude Opus 5.5 with lower costs
Anthropic launched Claude Opus 5.5 with Fable-level performance claims, lower operating costs and stronger safeguards for cybersecurity work.
What happened
Anthropic released Claude Opus 5.5 on September 22 as the first model in its Claude 5.5 line. The company says the model reaches roughly Claude Fable 5.1 performance on most tasks while costing 40% less to run than Opus 5. Anthropic also reports that Opus 5.5 generates answers more than 30% faster and uses fewer tokens on comparable work.
The model is available across Claude's consumer and business products, as well as through the API and major cloud distribution channels. It is already appearing in GitHub Copilot, Cursor, AWS and Google Cloud, giving coding and enterprise users immediate access.
Safety and deployment
Anthropic framed the launch around both capability and control. Opus 5.5 was tested by external evaluators including METR and Frontier Design, and Anthropic says it achieved the strongest result yet on its broadest alignment evaluation. The company also highlighted stricter safeguards for cybersecurity-related requests, an area where more capable coding models can be useful to defenders but dangerous in the wrong hands.
The timing is notable. Anthropic has recently argued for slowing the frontier, yet it is now shipping a model that claims better agentic coding and knowledge-work performance at materially lower cost. That tension makes efficiency part of the safety story: cheaper, faster models can increase beneficial use, but they can also make large-scale misuse easier to automate.
Why it matters
Opus 5.5 raises the competitive pressure on both model quality and unit economics. Its importance will depend less on Anthropic's headline benchmark claims than on whether the model maintains reliability over long coding sessions and whether its safety controls hold when connected to tools, repositories and cloud systems.