Anthropic Drops Claude Opus 4.6: The New King of Reasoning?
New 1M token context window and state-of-the-art performance on complex reasoning and coding benchmarks.
Deep Dive
Anthropic launched Claude Opus 4.6, its most advanced AI model. It features a 1M token context window in beta and outperforms OpenAI's GPT-5.2 by 144 Elo points on the GDPval-AA benchmark for professional tasks. The model excels at agentic coding, complex reasoning, and autonomous work in tools like Claude Code and Claude in Excel. It's available now via API and claude.ai at the same price as its predecessor.
Why It Matters
This directly challenges OpenAI's lead in high-stakes professional domains like finance, legal, and coding with measurable performance gains.