Claude Opus 4.8

Claude Opus 4.8 is Anthropic’s most intelligent Opus model, delivering advanced performance across coding, agentic workflows, and enterprise applications. It is designed for long-running autonomous execution, with improved tool use and the ability to manage complex, multi-step tasks while maintaining progress over extended periods with minimal oversight. In coding, it operates like a subject matter expert, understanding large codebases, planning before making changes, and handling long, multi-stage development tasks while tracking dependencies and resolving blockers independently. For enterprise workflows, it excels at instruction following, maintaining scope, and producing professional-quality outputs across documents, spreadsheets, and presentations, while reasoning across long contexts and multi-stage projects. It brings deeper reasoning and precision to financial analysis, handling dense filings, charts, and long reporting cycles with high fidelity. In cybersecurity, it supports advanced workflows such as threat analysis, vulnerability detection, and incident response by maintaining context over long traces and large systems. It also offers strong computer-use capabilities, combining improved vision with planning and reasoning to execute multi-step tasks across multiple applications.