Anthropic’s Claude Opus 4.7 has outperformed OpenAI’s new GPT-5.5 in a series of challenging reasoning and logic tests, despite GPT-5.5’s strong performance in agentic coding benchmarks. The ...