Independent benchmarking platform LM Arena updated its leaderboards on 18 Aug. 2025, showing Anthropic’s newly released Claude Opus 4.1 Thinking model debuting at No. 1 in the Text, Coding and WebDev arenas. It is the first system to hold the top spot across all three flagship categories simultaneously, according to the site’s publicly available rankings. Both the thinking and non-thinking versions of Claude Opus 4.1 now occupy first and second place respectively on the coding list, displacing OpenAI’s GPT-5-high. The update consolidates Claude’s position in a field where performance differentials can determine which models are selected by corporate developers and research teams.