Anthropic has launched Claude Opus 5.5, the first model in its new Claude 5.5 family. The company says the model delivers performance comparable to its higher-end Fable 5.1 model on most tasks while costing significantly less to operate.
The launch comes as AI companies continue to compete on coding, agentic tasks and professional work, with Anthropic putting particular emphasis on long-running software projects and improved efficiency.
Claude Opus 5.5 is built for complex work
Anthropic is positioning Opus 5.5 for tasks that require an AI system to work through multiple steps rather than simply answer individual questions.
Coding is one of the main areas of focus. Anthropic says early testers used the model for large codebase migrations, software audits and fixing performance problems across applications. In one test reported by the company, Opus 5.5 completed a 200,000-line codebase audit and fixes in under three hours, while Opus 5 took more than 20 hours.
The model is also designed for knowledge work and computer-use tasks, making it relevant to businesses using AI agents for longer workflows.
40% lower typical workload cost
One of the biggest changes is pricing.
Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens. The previous Opus 5 pricing was $5 per million input tokens and $25 per million output tokens.
Anthropic says typical workloads cost about 40% less to run than Opus 5. Cache reads, which can account for a large part of the cost of coding and agentic workloads, are also cheaper.
This could be particularly important for developers and businesses running AI agents for hours rather than using AI for occasional questions.
More than 30% faster than Opus 5
Anthropic says Opus 5.5 also generates output more than 30% faster than its predecessor.
The company attributes the combination of lower prices and higher speed to reduced computing requirements for serving the model. For users running long coding or research tasks, that can reduce both waiting time and overall operating costs.
Safety is a major part of the release
The launch also comes after a series of concerns around AI systems interacting with real-world computer systems during testing.
Anthropic says Opus 5.5 was externally evaluated by organizations including METR and Frontier Design. The company also says the model achieved its strongest result so far on its most comprehensive alignment test.
For higher-risk cybersecurity and biological requests, Anthropic has added additional safeguards. The company says some requests can be redirected to less capable models when its systems determine that additional restrictions are required.
Claude 5.5 family will expand
Opus 5.5 is only the beginning of the new Claude 5.5 generation.
Anthropic says Claude Sonnet 5.5 and Claude Haiku 5.5 are expected to arrive in the coming weeks. That would extend the same generation across different performance and cost levels.
Opus 5.5 is already available through Anthropic's services, and GitHub has also announced availability of the model in GitHub Copilot for eligible plans and products.
What Claude Opus 5.5 means for developers
For developers, the most significant change may not simply be the model's benchmark performance. The combination of stronger long-running coding capabilities, lower token prices and faster generation could make it more practical to use AI agents for larger software projects.
The arrival of Sonnet 5.5 and Haiku 5.5 will also be important because many developers choose models based on the balance between capability, speed and operating cost.
Comments
Post a Comment