Introducing Claude Opus 4.6 - We’re upgrading our smartest model.

Anthropic announced the release of Claude Opus 4.6, the latest update to its flagship large language model, on February 5, 2026. The new model introduces significant enhancements to its agentic coding capabilities, designed to handle larger codebases and sustain complex tasks for longer periods. A key development is the introduction of a 1 million token context window in beta, a first for the company's Opus-class models, which enables the processing of extensive documents and code repositories in a single pass. The update is aimed squarely at professional use cases where AI is required to operate with more autonomy and precision on complex, multi-step work.

According to the company's data, Opus 4.6 achieves state-of-the-art performance on several industry benchmarks. It reportedly secured the highest score on the agentic coding evaluation Terminal-Bench 2.0 and outperformed OpenAI’s GPT-5.2 by 144 Elo points on GDPval-AA, an evaluation of economically valuable tasks. The release also includes new developer-focused features, such as 'adaptive thinking' to modulate reasoning depth and 'effort controls' to balance intelligence against speed and cost. Despite the upgrades, Anthropic has kept the pricing consistent with the previous version at $5 per million input tokens and $25 per million output tokens, making the model available immediately through its API, consumer-facing website, and major cloud platforms.

The launch of Opus 4.6 suggests a market-wide push toward AI systems that function less like conversational assistants and more like autonomous agents for specialized labor. Early feedback from access partners highlights the model's ability to break down complex requests, manage subagents, and operate with less human oversight, particularly in software development and cybersecurity investigations. By integrating these advanced capabilities into enterprise tools like Excel and a new research preview of PowerPoint, Anthropic is positioning Claude as a practical collaborator embedded within existing professional workflows, competing on reliability and long-term task completion in addition to raw performance metrics.

Anthropic's release of Opus 4.6 signals a strategic focus on making AI agents not just more capable, but more reliable and economically viable for long-duration, complex professional tasks, directly challenging competitors on practical enterprise utility rather than just raw intelligence benchmarks.