Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price.

Claude for Excel, PowerPoint, and Word are now generally available, and Claude for Outlook is in public beta. As Claude moves between your Microsoft apps, it carries the full context of your conversation.

Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.

Claude for Word is now in beta. Draft, edit, and revise documents directly from the sidebar. Claude preserves your formatting, and edits appear as tracked changes. Available on Team and Enterprise plans.

Claude Cowork is now generally available to all paid plans. For Enterprise, we are adding role-based access controls, group spend limits, usage analytics, and expanded OpenTelemetry to give admins what they need to deploy it across the org.

We're bringing the Claude in Office ecosystem closer together. Would love to know what you think. I poured my soul into this one and want it to feel great!

I’m so grateful to work here. Very proud of the tough, principled decisions we’re making here.

This is Claude Sonnet 4.6: our most capable Sonnet model yet. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also features a 1M token context window in beta.

Simile is out of stealth! At Simile, we have built the first AI simulation of society, populated by agents based on real humans. We are building a foundation model that predicts human behavior in any situation, and a product that deploys it at scale. Thrilled to be on this mission.

Truly privileged to have worked on this! We're bringing Claude to millions of knowledge workers - beyond just coding. If you're heavy in finance (or just an Excel nerd), let us know what you think.

Finally, AI designs you can edit. Moda - create brand assets on a fully editable canvas. This is not another ChatGPT or nano banana wrapper. We actually taught an agent layout, typography, and color. We gave it taste. Use it for slides, social, infographics, e-books …

We spent the first year at @GentraceAI helping customers run traditional LLM evaluations. But we realized late last year that agents changed everything. Agents don't just generate text - they take actions, make decisions, and interact across multiple steps. Traditional "hallucination" checks can't evaluate whether your agent booked the right flight or resolved the customer’s support ticket correctly. Old LLM observability tools were built for simple input/output prompts. They break in an agentic world. For the last 9 months, we rebuilt Gentrace from first principles 🔥 Here's our new approach: - Chat with your AI trace data using our agent to discover what's actually breaking - Describe problems in plain English - Our agent creates custom AI-powered monitoring columns that catch these failures across all future traces Example: Instead of checking "factual accuracy," you can now evaluate complex behaviors: - "Figure out if a user is frustrated with my AI agent in the customer support chat" - "Detect when my agent takes more than 5 steps to answer simple queries." We stopped measuring outputs and started measuring outcomes. Watch @dougsafreno break down how to solve agent observability with Gentrace. 👇

Just released a 20 minute beginner tutorial making a flight tracker app with @alexcodes_ai Check it out if you're just getting started!

This team's progress blows me away every time I see an update.

Grep.app works incredibly well but desperately needs a "Copy files to Markdown" button. cc @rauchg