Anthropic has released Claude Opus 4.7, scoring 64.3% on SWE-bench Pro — ahead of GPT-5.4 (57.7%) and Gemini 3.1 Pro (54.2%). Key improvements include a 14% gain in multi-step agentic reasoning with a third of the tool errors, multi-agent coordination for parallel workstreams, 3x higher image resolution (up to 2,576px), and
Table of contents
Where it leadsThe agentic stepVision and contextPricing and availabilityWhat it meansSort: