Anthropic has released Claude Opus 4.6, an upgrade to its smartest model that enhances coding skills, agentic task sustainability, and code review capabilities. The update introduces a 1M token context window in beta and improves performance on complex reasoning and financial analysis tasks.
- Achieves state-of-the-art scores on Terminal-Bench 2.0, Humanity’s Last Exam, and BrowseComp.
- Outperforms GPT-5.2 by approximately 144 Elo points and its predecessor Claude Opus 4.5 by 190 points on GDPval-AA.
- Introduces adaptive thinking and effort controls in the API to manage intelligence, speed, and cost.
- Expands into PowerPoint via research preview and enhances capabilities within Claude in Excel.
The model is available today on claude.ai, the API, and major cloud platforms with pricing remaining at $5/$25 per million tokens.