Anthropic has officially rolled out Claude Sonnet 5.5, an advanced iteration of its Claude Sonnet 5 model. This new AI is engineered for enhanced speed, improved efficiency, and superior performance across a range of tasks, particularly coding and other knowledge-intensive work.
Designed to streamline everyday operations, such as document creation and complex problem-solving, Claude Sonnet 5.5 achieves these improvements while demanding fewer computational resources. Anthropic reports that the model generates responses more than 30% faster than its predecessor. Furthermore, despite maintaining the same token pricing, it can reduce the cost per task by as much as 30% due to greater efficiency.
Significant Boost in Coding and Collaboration
One of the most notable advancements in Claude Sonnet 5.5 is its strengthened coding prowess. The model achieved an impressive 70.6% on Terminal-Bench 4.0, a substantial leap from Sonnet 5's 10.3%. It also scored 55.5% on CursorBench 4.0, demonstrating improvements in handling long-horizon knowledge work and image understanding. Anthropic proudly states that Sonnet 5.5 is the first Sonnet model capable of beating Pokémon Red solely by interpreting screenshots.
Beyond coding, the new model offers enhanced collaboration features. It is designed to write more clearly and is better suited for rapid iterative processes. Early testers have highlighted its ability to quickly comprehend complex codebases, requiring fewer steps and tool calls to complete tasks.
Cost-Efficiency Through Optimization
While Claude Sonnet 5.5 retains the same token pricing as its predecessor—$2 per million input tokens, $10 per million output tokens, and $0.20 per million cache-read tokens—its core advantage lies in efficiency. By utilizing fewer tokens to accomplish the same tasks, the model effectively lowers the overall cost per task by up to 30% for users.
Anthropic positions Sonnet 5.5 as a faster and more economical alternative to its premium Claude Opus 5.5 model, particularly for well-defined tasks that do not necessitate the continuous nuanced judgment characteristic of its top-tier offering. This version also marks the first Sonnet model to integrate enhanced cybersecurity measures and fallback options.
Availability and Access
Claude Sonnet 5.5 is now accessible across all of Anthropic's supported platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. Developers can integrate and utilize the model through the Claude Platform by specifying the model identifier claude-sonnet-5-5.