Anthropic Releases Claude Opus 4.7 with Enhanced Agentic Coding and Vision Capabilities

Anthropic launched Claude Opus 4.7, achieving 98.5% visual acuity (up from 54.5%), 3x resolution on image inputs, and 21% fewer document reasoning errors. The model leads benchmarks on SWE-bench and agentic reasoning, outperforming GPT-5.4 and Gemini 3.1 Pro while maintaining the same $5/$25 per million token pricing. The more powerful Mythos model remains restricted to enterprise security partners.