An update on recent Claude Code quality reports
…We’ve traced these reports to three separate changes that affected Claude Code, the Claude Agent SDK, and Claude Cowork. The API was not impacted. All three issues have now been resolved…
Search
Natural Language Autoencoders
…Related content Teaching Claude why New research on how we've reduced agentic misalignment. Donating our open-source alignment tool Focus areas for The Anthropic Institute At The Anthropic Institute (TAI), we…
Building AI for cyber defenders
…Evaluating AI Agents' Cybersecurity Capabilities with Real-World Vulnerabilities at Scale," arXiv preprint arXiv:2506.02548 (2025), https://arxiv.org/abs/2506.02548 . Related content Agentic coding and persistent returns to expertise…
Anthropic raises $65B in Series H funding at $965B post-money valuation
…Related content Introducing Claude Opus 4.8 An upgrade to our Opus class of models, with stronger performance across coding, agentic tasks, and professional work, and the consistency to handle long-running…
Anthropic Sydney office
…partnership will bring Claude's AI directly into Xero —and Xero's financial data and tools into Claude.ai. We're also working with YMCA South Australia as a Claude for Nonprofits…
Evaluating Claude’s bioinformatics research capabilities with BioMysteryBench
…Although these benchmarks were developed in the “chatbot” era, they’ve persisted into the agent and tool-use era, joined by even more difficult scientific reasoning evals like FrontierScience and Humanity's…
Anthropic opens Seoul office and announces new partnerships across the Korean AI ecosystem
…Thousands of NAVER engineers are now using Claude Code to diversify their coding tools and maximize coding productivity. At global online game company Nexon, engineering teams use Claude Code to write, review…
How AI Is Transforming Work at Anthropic
…I use the tools for both things that are core to my expertise (as an accelerant, where I know what to expect and can guide the agent effectively), and for things that…
More details on Fable 5’s cyber safeguards and our jailbreak framework
…1 Somewhat better than existing tools; can help a novice make partial progress on an offensive task beyond existing tools. Not useful for domain experts. 2 Hard or costly to obtain with…
Project Glasswing: An initial update
…The thoughtful use of publicly available AI models can help here; we’re building tools and sharing our research to support this (more details below). Developers should also help their users stay…