Anthropic unveiled Claude Opus 4.6 on Thursday, introducing "agent teams" that enable parallel task execution and positioning the model as a versatile tool for knowledge workers beyond software development.

The standout feature allows multiple agents to split complex tasks into segmented jobs and coordinate simultaneously. "Instead of one agent working through tasks sequentially, you can split the work across multiple agents—each owning its piece and coordinating directly with the others," Anthropic explains. Scott White, Head of Product, compared it to managing a talented human team working in parallel for faster results. Agent teams are available in research preview for API users and subscribers.

Opus 4.6 now offers a 1 million token context window—matching Sonnet 4 and 4.5—enabling work with larger codebases and documents. The model also achieves state-of-the-art performance on several evaluations, including Terminal-Bench 2.0 for agentic coding and Humanity's Last Exam for complex reasoning. On GDPval-AA, which measures economically valuable knowledge work, Opus 4.6 outperforms OpenAI's GPT-5.2 by approximately 144 Elo points.

PowerPoint integration represents another significant upgrade. Previously, users created presentations in Claude and transferred them to PowerPoint for editing. Now Claude works directly within PowerPoint as an accessible side panel, streamlining the workflow.

According to White, Opus has evolved from a specialized software development tool into a platform serving product managers, financial analysts, and professionals across industries. "We noticed a lot of people who are not professional software developers using Claude Code simply because it was a really amazing engine to do tasks," he told TechCrunch.