Claude Opus 4.6

Anthropic announced the release of Claude Opus 4.6, an update to its most capable large language model, focused on enhancing its performance in coding and complex, multi-step tasks. The new model demonstrates state-of-the-art results on several key industry benchmarks, including evaluations for agentic coding and multidisciplinary reasoning. Notably, Opus 4.6 is being introduced with a one-million-token context window in beta, a significant expansion aimed at improving performance on tasks requiring the synthesis of large amounts of information.

According to the company's data, Opus 4.6 achieves the highest scores on agentic coding evaluation Terminal-Bench 2.0 and outperforms the next-best model, OpenAI’s GPT-5.2, by 144 Elo points on GDPval-AA, a benchmark for economically valuable knowledge work. The model is available immediately through Anthropic's API and claude.ai, with pricing remaining unchanged at $5 per million input tokens and $25 per million output tokens. Anthropic also introduced new developer controls, including 'adaptive thinking' to adjust compute use based on task complexity and an 'effort' parameter to manage the trade-off between intelligence, speed, and cost.

This release positions Anthropic to compete more directly in the enterprise and developer markets, where demand is growing for AI that can reliably execute long-running, autonomous tasks within complex software environments. By improving long-context retrieval and addressing common issues like 'context rot,' Opus 4.6 is engineered for practical professional applications such as financial analysis, legal reasoning, and managing large codebases. The focus on measurable performance in specialized, high-value domains indicates a market shift towards models judged not just on general capabilities but on their ability to be integrated into specific, critical business workflows.

Strategic Takeaway: Anthropic's strategy with Opus 4.6 is a direct play for the enterprise developer market, prioritizing demonstrable performance on complex, long-horizon tasks and professional workflows. By improving long-context reliability and agentic planning, the company aims to position Claude as the go-to platform for building autonomous systems that handle economically valuable work.