Anthropic: Claude Opus 4

Anthropic's Claude Opus 4, released May 22, 2025, demonstrates exceptional reliability with a 100% success rate across all benchmarks, indicating consistent and usable responses. While its speed performance is moderate, ranking in the 35th percentile, its pricing tends to be premium, positioned at the 3rd percentile. The model excels in several key areas, achieving perfect accuracy in Hallucinations (100%), General Knowledge (100%), and Ethics (100%), often being the most accurate model at its price point and speed for these categories. It also shows strong performance in Coding (94% accuracy, 86th percentile), Mathematics (94% accuracy, 77th percentile), and Reasoning (90% accuracy, 72nd percentile), aligning with its description as a leading coding model. Its instruction following capabilities are solid at 71% accuracy (76th percentile). A relative weakness appears in Email Classification, where its 97% accuracy places it in the 38th percentile, suggesting room for improvement compared to its other high-performing categories. Overall, Claude Opus 4 is a highly reliable and accurate model, particularly strong in complex cognitive tasks and coding, albeit at a premium price point and moderate speed.