Anthropic: Claude Sonnet 4.5

Anthropic's Claude Sonnet 4.5, created on September 29, 2025, demonstrates a strong performance profile, particularly excelling in reliability with a 100% success rate across all benchmarks. While its speed performance is moderate, ranking in the 33rd percentile, its pricing tends to be at a premium level, positioned in the 11th percentile. The model exhibits exceptional accuracy in critical areas. It achieved perfect scores in Hallucinations (100%), General Knowledge (100%), and Ethics (100%), often being the most accurate model at its price point and speed for these categories. Sonnet 4.5 also shows state-of-the-art capabilities in Coding (95.0% accuracy, 94th percentile) and strong performance in Mathematics (94.0% accuracy, 86th percentile). Its agentic capabilities are further highlighted by solid results in Instruction Following (67.7% accuracy, 77th percentile) and Reasoning (88.0% accuracy, 77th percentile). Email Classification is also robust at 99.0% accuracy. The model's design for extended autonomous operation, improved tool orchestration, and efficient context management positions it well for complex, long-running workflows in domains like software engineering and financial analysis.