Anthropic: Claude Opus 4.8

Anthropic's Claude Opus 4.8, created on May 27, 2026, is positioned as a highly capable model within the Opus family, supporting multimodal inputs and a 1M-token context length. In terms of speed, it demonstrates competitive response times, performing around the 57th percentile across various benchmarks. However, its pricing tends to be at premium levels, ranking in the 9th percentile for cost-effectiveness. A significant strength is its exceptional reliability, boasting a 99% success rate, indicating minimal technical failures. The model exhibits strong performance across several critical benchmarks. It achieves perfect accuracy in Ethics, making it the most accurate model at its price point and among models of similar speed. It also excels in Mathematics (86th percentile) and Reasoning (75th percentile), showcasing robust analytical capabilities. Coding performance is solid at the 70th percentile, and General Knowledge is nearly perfect at 99.5% accuracy. While its Instruction Following (67th percentile) is good, its Hallucinations score (44th percentile) suggests some room for improvement in acknowledging uncertainty, though 94% accuracy is still commendable. Email Classification is average at the 49th percentile. Overall, Claude Opus 4.8 is a highly reliable and accurate model, particularly strong in ethical reasoning, mathematics, and complex problem-solving, albeit at a premium price point.