Opus 4.6 is Anthropic's most intelligent model to date, designed for coding, complex agent workflows, and high-stakes enterprise tasks. It represents an upgrade from Opus 4.5, featuring improvements in reasoning, instruction-following, and overall performance. Anthropic highlights its ability to handle more complex tasks and deliver faster results. Opus 4.6 is particularly strong in coding, demonstrating the ability to deliver production-ready code with minimal oversight, operate reliably in larger codebases, and review/debug code effectively. It excels in tasks such as financial analysis, research, and working with documents, spreadsheets, and presentations.
Key features of Opus 4.6 include a 1 million token context window (in beta), allowing it to ingest and process vast amounts of information. It also supports outputs of up to 128,000 tokens. The model introduces an 'adaptive thinking' mode, replacing the previous 'extended thinking' mode, which allows the model to dynamically adjust its reasoning depth. Opus 4.6 is available through the Claude API and on platforms like Amazon Bedrock, Google Cloud's Vertex AI, and Microsoft Foundry. It is also integrated into tools like Microsoft Copilot Studio, Kiro IDE and CLI, and Windsurf.
Pricing for Opus 4.6 starts at $5 per million input tokens and $25 per million output tokens for the standard context window. For developers using the 1 million token context window (beta), the pricing shifts to a premium rate. The cost is $10 per million input tokens and $37.50 per million output tokens for requests exceeding the standard 200K token threshold. While the base cost is the same as Opus 4.5, new features may come with a premium.
