Anthropic’s new Claude Opus 4.5 surpasses GPT-5.1 and Gemini 3 in coding and agent benchmarks, while offering improved safety and real-world task performance.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results