Anthropic Unveils Claude Opus 4.6 with 1M Token Window

Anthropic on Thursday unveiled Claude Opus 4.6, the latest upgrade to its flagship artificial intelligence model, introducing a one-million-token context window and multi-agent collaboration features designed to handle complex tasks more efficiently.

The new version builds on Claude Opus 4.5, released last November, by expanding the model's capacity to process vast amounts of information in a single interaction. Previously limited to 200,000 tokens, equivalent to several hundred pages of text, Opus 4.6 now supports up to one million tokens in beta on the Claude Developer Platform. This allows it to manage inputs such as 1,500 pages of documents, 30,000 lines of code or more than an hour of video without losing earlier details. For extended conversations or coding sessions, the system compresses prior exchanges to preserve the context window's limits. Anthropic noted that this capability aligns Opus 4.6 with competitors like Google's Gemini 1.5 Pro, which also handles one million tokens.

A key addition is Agent Teams, a feature that enables multiple AI agents to divide and tackle large projects in parallel. Rather than a single agent processing tasks sequentially, users can...