Anthropic: Claude Sonnet 4

Anthropic's Claude Sonnet 4, released May 22, 2025, demonstrates a strong balance of capability and efficiency, particularly excelling in coding and reasoning tasks. It achieves state-of-the-art performance on SWE-bench (72.7%) and boasts exceptional reliability with a 100% success rate across all benchmarks, indicating consistent and usable responses. While its speed performance is competitive, ranking in the 41st percentile, its pricing tends to be at premium levels (11th percentile). Key strengths include perfect accuracy in Hallucinations and Ethics benchmarks, showcasing robust truthfulness and adherence to ethical principles. It also exhibits high accuracy in Coding (93.0%), General Knowledge (99.8%), and Mathematics (93.0%), indicating strong foundational and advanced problem-solving abilities. Its Reasoning capabilities are also notable at 88.0% accuracy. A potential area for improvement is Instruction Following, where it achieved 66.0% accuracy, though still performing above average. Sonnet 4 is optimized for practical everyday use, offering advanced reasoning and reduced error rates in agent-driven workflows.