Anthropic has launched Claude Opus 5.5, the first in a new model series. The company reports that the model matches the performance of Claude Fable 5.1 on most tasks while operating at a 40 percent lower total cost compared to its predecessor, Opus 5.
The pricing for Opus 5.5 is set at $4 per million input tokens and $20 per million output tokens, representing a 20 percent reduction from previous rates. Additionally, cache read costs have been decreased by 60 percent, and the model generates output more than 30 percent faster.
In benchmark testing, Anthropic claims Opus 5.5 outperforms OpenAI's GPT-6 Astra on the FrontierCode test at approximately 20 percent of the cost per task. On Terminal-Bench 4.0, it matches Astra's performance at 40 percent of the cost, and on CursorBench, it exceeds GPT-5.6 Sol by 11 points at one-third of the cost.
The new model features improved communication style, aiming to reduce the formulaic writing previously criticized by users. It also incorporates safety safeguards for cybersecurity and biology that align with Fable 5.1 standards, with specific tasks routed to older models when necessary.
Opus 5.5 is available across major cloud platforms including Amazon Web Services, Google Cloud, and Microsoft Azure. The release includes anti-distillation measures such as Preserved Thinking and watermarking to comply with the EU AI Act, alongside a Zero Data Retention option.
