Anthropic's Claude Opus 4.7 has edged out OpenAI's ChatGPT-5.5 in a head-to-head comparison of complex reasoning tasks. In multi-step probability problems, Claude not only reached the correct answer ...