AsianFin -- Anthropic on Monday released Claude Sonnet 4.5, describing it as “the best coding model in the world,” alongside a suite of new developer tools. The company highlighted the model’s ability to maintain focus for over 30 hours on complex, multi-step coding tasks, while demonstrating improvements in reasoning and mathematical capabilities.

Claude Sonnet 4.5 scored 77.2% on SWE-bench Verified, a benchmark measuring real-world software coding skills. With parallel test-time compute, the score rises to 82%, placing it ahead of comparable models from OpenAI and Google, as well as Anthropic’s own Claude 4.1 Opus.

Anthropic’s naming scheme categorizes Haiku as a small model, Sonnet as medium, and Opus as the heaviest and most powerful in the family, underscoring the significance of Sonnet 4.5’s performance gains.