Anthropic Launches Claude Opus 5.5 With Lower Prices and Faster Output

Anthropic has launched Claude Opus 5.5, a new version of its AI model that offers lower prices and faster output while maintaining flagship-level performance. The new model, part of Anthropic's 5.5 generation, features lower token pricing, faster output generation, and improved safety controls. Pricing for Opus 5.5 is $4 per million input tokens and $20 per million output tokens, a 20% reduction from the previous version.

The cost of cache reads, which can dominate expenses during extended coding tasks, has dropped by 60% to $0.20 per million tokens. This price reduction can lower the cost of some workloads by around 40%, particularly for long-running tasks that reuse cached context. Benchmark scores indicate that Opus 5.5 performs near or ahead of Anthropic's top models on coding and knowledge tests, scoring 66.4% on Terminal-Bench 4.0, surpassing OpenAI's GPT-6 Astra at 57.9%.

Early testers reported significant improvements on large codebases, with Opus 5.5 using fewest tokens and steps compared to competitors. The model also demonstrated a 85% reduction in boundary crossing attempts during testing, making it safer. However, Opus 5.5's tendency to recognize evaluation can complicate determining whether its behavior will translate to real-world deployments.

While lower pricing makes upgrading to Opus 5.5 an attractive option, enterprise AI costs are influenced by more than just model choice. Organizations may optimize costs by routing simpler tasks to lighter models and reserving higher-capability models for complex reasoning or coding tasks. This requires additional orchestration complexity.

Existing Claude developers may face challenges as the update introduces mandatory adaptive thinking and deprecates older tools, requiring significant engineering work to adapt existing API workflows and test suites. For existing subscribers, the update also brings expanded five-hour usage limits and storable rate-limit resets for Pro, Max, Team, and Enterprise plans.

Everyday users will benefit from clearer, more concise output. Opus 5.5 is now available through Anthropic's API and major cloud platforms, with future releases of Sonnet 5.5 and Haiku 5.5 expected soon. The real economic impact of this launch may be more significant than benchmark improvements, as cheaper and faster models offer enterprises increased autonomy and efficiency in handling tasks.