Claude

When Claude gets something wrong, it could be a model or an effort problem. To figure that out, ask yourself: did Claude not try hard enough, or did it not know enough? Opus at low effort is like getting five minutes with an expert who has deep experience with problems like yours. Sonnet at high effort is like giving a really good generalist the whole afternoon. Fable, even at low effort, is that specialist glancing at the problem everyone else is stuck on and still spotting the thing no one else would. None of these is universally better. Most real tasks need some of both.

Introducing Claude Opus 5: a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price. On several coding and knowledge work evaluations, Opus 5 is the new state-of-the-art. It’s also highly efficient—it outperforms other models for a similar or lower cost per task. According to our automated behavioral audit, Opus 5 is our most aligned model to date. It shows the lowest rates of reckless or deceptive behavior, and the strongest adherence to Claude’s Constitution. It’s available today on all paid plans and the Claude API, priced the same as Opus 4.8. It’s the default model on Claude Max, and the strongest on Claude Pro. Opus 5 is also available in Fast mode, which runs around 2.5× the default speed. Read more: https://lnkd.in/eA5EsPHt

Voice conversations with Claude now use more of the models you have in chat. Talk through real strategy questions and judgment calls out loud, and switch models right in the conversation as the problem gets harder. When you're ready to act, Claude reaches the tools you've connected. Ask it to catch you up on the morning's email, or push your next meeting back by 30 minutes. Voice mode also supports more languages, on all plans, so you can talk in the language you think in. The update is rolling out today on mobile, desktop, and web. Voice mode is still in beta and works best from your phone. Download the Claude app and tap the sound wave symbol to start a conversation: claude.com/download

The Claude Security plugin for Claude Code is now available in beta. You can now scan your code for vulnerabilities before you commit, or run a full scan across your codebase, right from the terminal. Once installed, /claude-security runs on the Claude inference you already run: - Scan codebase: Maps your repository's architecture, creates a threat-model, and sends researcher agents across the code. - Scan changes: Points the engine at a pull request, a single commit, or the changes on a branch. - Suggest patches: Turns confirmed findings into patches, with an adversarial panel reviewing each one before it reaches you. Read more: https://lnkd.in/eycXybGk

You can now ask Claude how people in your field are using AI. The answers come from the Anthropic Economic Index, our public dataset measuring how AI is used across the economy: which occupations use it most, and for what kinds of tasks. Researchers, journalists, and policymakers have used it to follow those patterns. Now it's a connector in claude.ai, so anyone can explore the data by asking questions. Ask the way you'd ask a colleague: • "Which occupations use AI the most?" • "What do people in education use Claude for?" • "How do people in Colorado use it?" To get started, open the connectors menu in claude.ai and add the Anthropic Economic Index. It works in any conversation, and the full datasets remain free to download: https://lnkd.in/eCi-9h2x

We spoke with Yan Liu about the ideas she used to talk herself out of building. During her month of experiments in Claude Design, she made a match-3 game, a liquid glass refraction, a 3D tulip atlas, and a receipt covered in passport stamps.

New in Claude Cowork: teach Claude a skill. Record your screen while you do a task, talking through it as you go. Claude turns the recording into a skill and saves it, so the next time that task comes up, you run the skill instead of repeating the steps. To try it, open Cowork in the Claude desktop app and choose Record a skill from the + menu. Available on Pro, Max, and Team plans.

When a used milling machine arrived at BLA Studios with a manual written in German, Joe Bacus used Claude to translate and troubleshoot it until the machine was running. It's the kind of problem Joe and Leef Armontrout take on constantly at their two-person fabrication studio in St. Louis: one they've never solved before. Read more: https://lnkd.in/dma2ZFXV