Claude Opus 4.6: A New Era of AI is Here

Anthropic has significantly upgraded its most powerful model, Claude Opus 4.6, delivering substantial improvements across a range of capabilities. This isn’t just an incremental update; it represents a leap forward in AI performance, particularly in coding, complex reasoning, and safety. As the AI landscape evolves with models like Google Gemini, Claude Opus 4.6 stakes its claim as a leading contender.

Enhanced Coding Prowess

Opus 4.6 demonstrates markedly improved coding skills. It exhibits more careful planning, sustains agentic tasks for longer durations, operates more reliably within larger codebases, and boasts superior code review and debugging capabilities – even identifying and correcting its own errors. This makes it an invaluable tool for developers seeking to streamline their workflow and enhance code quality.

A Massive 1 Million Token Context Window

In a groundbreaking first for Opus-class models, Claude 4.6 introduces a 1 million token context window (currently in beta on the Claude Developer Platform). This expanded context allows the model to process and understand significantly larger amounts of information, leading to more nuanced and accurate responses. This is a game-changer for tasks requiring deep analysis of extensive documents or complex datasets.

Beyond Coding: Versatility for Everyday Tasks

The improvements in Opus 4.6 extend beyond coding. It excels at everyday work tasks such as financial analysis, research, and document creation (spreadsheets, presentations, and more). Within Cowork, Claude’s multitasking environment, Opus 4.6 can autonomously handle complex projects, leveraging its enhanced skills to deliver exceptional results.

Benchmark Performance: Leading the Pack

Opus 4.6 consistently achieves state-of-the-art performance on several key evaluations:

• Terminal-Bench 2.0: Achieves the highest score on this agentic coding evaluation.
• Humanity’s Last Exam: Leads all other frontier models on this complex multidisciplinary reasoning test.
• GDPval-AA: Outperforms OpenAI’s GPT-5.2 by approximately 144 Elo points and Claude Opus 4.5 by 190 points on economically valuable knowledge work tasks.
• BrowseComp: Demonstrates superior ability to locate hard-to-find information online.

Safety First: A Commitment to Responsible AI

Anthropic prioritizes safety. Opus 4.6 maintains an excellent safety profile, comparable to or better than other leading models. It exhibits low rates of misaligned behavior across various safety evaluations. The company has implemented new safeguards, particularly in areas like cybersecurity, to mitigate potential misuse.

New Features and Updates

• Agent Teams in Claude Code: Assemble teams of agents to work collaboratively on tasks.
• Context Compaction: Summarize context to enable longer-running tasks without exceeding limits.
• Adaptive Thinking: Dynamically adjust the level of reasoning based on contextual clues.
• Effort Controls: Fine-tune intelligence, speed, and cost.
• Claude in Excel: Improved performance on complex tasks, including data ingestion and multi-step changes.
• Claude in PowerPoint (Research Preview): Create visually appealing presentations from data and descriptions.

Availability and Pricing

Claude Opus 4.6 is now available on claude.ai, the Claude API, and major cloud platforms. Pricing remains consistent at $5/$25 per million tokens. Developers can access Opus 4.6 via the Claude API.

Built with Claude: A Cycle of Improvement

Anthropic’s engineers utilize Claude Code daily, ensuring that each new model is rigorously tested on real-world tasks. Opus 4.6 demonstrates a heightened focus on challenging aspects of a task, efficient handling of simpler elements, improved judgment in ambiguous situations, and sustained productivity over extended sessions.

The Future of AI is Here

Claude Opus 4.6 represents a significant advancement in AI technology, offering enhanced capabilities, improved safety, and a commitment to responsible development. As the competition heats up with models like Google Gemini, Anthropic continues to push the boundaries of what’s possible with AI.