Claude Sonnet 4.5 is a new large language model from Anthropic. It shows particular strengths in software coding, in “agentic” tasks where it runs in a loop and uses tools, and in using computers. It has substantial capability improvements compared to previous Anthropic models on evaluations in areas such as reasoning and mathematics.
This system card provides a detailed set of results from our evaluations of the model’s safety-related characteristics. It includes assessments related to model safeguards, agentic