Anthropic has announced Claude Opus 4.6, an upgraded AI model designed to handle complex reasoning, coding, and professional workflows with improved autonomy. The system plans tasks more carefully, reviews and debugs code, and operates reliably across large codebases. It also introduces a beta 1-million-token context window, allowing the model to process extensive information in a single session.
The company said the model can assist in financial analysis, research, and document production, including spreadsheets and presentations, while performing long-running agentic tasks. Benchmark testing showed strong results across multiple evaluations, including leadership on Terminal-Bench 2.0 and Humanity’s Last Exam, and higher performance on economically relevant knowledge-work tests than competing models. Pricing remains $5 input and $25 output per million tokens.
Industry partners highlighted practical gains. Sarah Sachs, AI Lead at Notion, said, “It takes complicated requests and actually follows through, breaking them into concrete steps, executing and producing polished work.” Mario Rodriguez, Chief Product Officer at GitHub, added, “Early testing shows Claude Opus 4.6 delivering on the complex, multi-step coding work developers face every day — especially agentic workflows that demand planning and tool calling.” The release also introduces adaptive thinking controls, context compaction for longer tasks, multi-agent teams, and deeper office-software integration.