Anthropic released Claude Opus 4.6 on February 5, bringing two significant enterprise capabilities: a 1 million token context window and agent teams for parallel task execution.
The context window expansion matches Sonnet 4 and 4.5, supporting analysis of large codebases, regulatory filings, and multi-document workflows in a single session. The agent teams feature, available in Claude Code (research preview for API users and subscribers), enables multiple AI agents to autonomously divide and coordinate on complex tasks like codebase audits or financial research.
Anthroplic positions this as mimicking human software team dynamics, letting agents work in parallel rather than sequentially. The company claims Opus 4.6 outperforms GPT-5.2 and Gemini 3 Pro on agentic planning, tool calling, and long-horizon tasks. Early partner feedback highlights autonomous execution on multi-step workflows.
Additional API features include compaction for sustained long-running tasks, adaptive thinking modes, and /effort controls to balance intelligence, speed, and cost. Anthropic also integrated Claude into PowerPoint via a side panel, targeting knowledge workers beyond developers.
The company reports 300,000-plus business users. This marks rapid iteration: Opus 4.5 shipped in November 2025, three months before this release.
What this means in practice: Enterprises running document-heavy workflows (legal review, financial analysis, technical due diligence) now have materially larger working memory. Agent teams could accelerate multi-stage processes like migration planning or security audits, though real-world validation will determine if the coordination overhead justifies the parallelization gains.
Worth noting: Anthropic's benchmarks show stronger performance than competitors, but benchmark-to-production translation remains the test. The claim of finding 500-plus zero-days in open source code (mentioned in social media reaction) would be significant if verified, though details weren't included in official materials.
The PowerPoint integration and agent teams suggest Anthropic is expanding beyond developer tools into broader enterprise productivity workflows. We'll see if enterprise buyers adopt multi-agent orchestration or if simpler single-agent implementations prove sufficient for most use cases.