Claude Opus 4.6

Claude Opus 4.6 is Anthropic's most intelligent model for building agents and coding. Significantly improved coding skills: more thorough planning, sustained agentic task support, reliable performance in large codebases, improved code review and debugging. Context window: 200K tokens by default, 1M tokens available in beta with premium pricing ($10/$37.50 per million input/output tokens at >200K). Up to 128K output tokens. New API features: adaptive thinking (the model decides when to use extended thinking), effort control (low/medium/high/max), context compression for long-running tasks. Leads on Terminal-Bench 2.0 (agentic coding), Humanity's Last Exam (multidisciplinary reasoning), GDPval-AA (knowledge work in finance, law), BrowseComp (information retrieval), DeepSearchQA (deep agentic search). Supports agent teams in Claude Code, Claude in Excel, and Claude in PowerPoint.