Claude Opus 4.6 is Claude's most capable model yet, built for deep reasoning, long-running agentic tasks, and large codebases. The model improves on its predecessor's coding skills with more careful planning, sustained agentic task performance, better reliability in larger codebases, and enhanced code review and debugging capabilities to catch its own mistakes. It also features a 1M token context window in beta, marking a first for Opus-class models.

The model demonstrates state-of-the-art performance across multiple domains including agentic coding, computer use, tool use, search, and finance. Key capabilities include improved long-context retrieval with less information drift, better reasoning after absorbing large amounts of information, and enhanced expert-level reasoning abilities. It can apply its improved abilities to everyday work tasks such as running financial analyses, conducting research, and using/creating documents, spreadsheets, and presentations.

Opus 4.6 operates with adaptive thinking capabilities where the model can pick up on contextual clues about how much to use its extended thinking. It features new effort controls giving developers more control over intelligence, speed, and cost. The model can perform context compaction to summarize its own context and handle longer-running tasks without hitting limits. Within Cowork, Claude can multitask autonomously, putting all these skills to work on behalf of users.

The model delivers substantial benefits across professional domains including coding, analysis, research, and real-world work tasks. Use cases span complex software engineering, cybersecurity investigations, legal reasoning, financial analysis, and multi-source analysis across technical, legal, and financial content. It excels at handling large codebase migrations, debugging unfamiliar codebases, and performing root cause analysis for complex software failures.

Target users include developers, knowledge workers, enterprises, and professionals across various domains who need advanced AI capabilities for complex tasks. The model integrates with Claude Code for agent teams, Claude Developer Platform APIs, and office tools including Claude in Excel and Claude in PowerPoint. Technical details include 1M token context window support, 128k output tokens, US-only inference options, and availability across major cloud platforms.