https://www.anthropic.com/anthropic 2026-09-23T13:49:17.335825+00:00 Anthropic python-feedgen https://www.anthropic.com/favicon.ico https://www.anthropic.com/favicon.ico Anthropic Newsroom, Research, Engineering, Red, Alignment Science, and Interpretability posts in one feed. https://www.anthropic.com/news/accenture-embedded-evaluation 2026-09-18T00:00:00+00:00 We’re partnering with Accenture on independent evaluation of frontier AI—part of our recent commitment to embed evaluators at Anthropic. Both we and Accenture expect to invest at least $1 billion to build capacity in this area over the next five years. 2026-09-18T00:00:00+00:00 https://www.anthropic.com/research/claude-uplifts-biomolecular-modeling 2026-09-17T00:00:00+00:00 How Claude is uplifting biomolecular modeling 2026-09-17T00:00:00+00:00 https://www.anthropic.com/news/life-sciences-verification-program 2026-09-17T00:00:00+00:00 Introducing the Life Sciences Verification Program 2026-09-17T00:00:00+00:00 https://www.anthropic.com/research/intelligence-targeting-conventional-weapons-capabilities 2026-09-10T00:00:00+00:00 Anthropic’s Frontier Red Team developed new evaluations to measure AI capabilities in tactical intelligence targeting and conventional weapons development. 2026-09-10T00:00:00+00:00 https://www.anthropic.com/research/alignment-assessment-cybersecurity-incidents 2026-09-09T00:00:00+00:00 We present an alignment assessment of four incidents in which Claude models gained unauthorized access to real third-party systems. 2026-09-09T00:00:00+00:00 https://www.anthropic.com/research/formalizing-fermats-last-theorem 2026-09-04T00:00:00+00:00 Formalizing Fermat's Last Theorem 2026-09-04T00:00:00+00:00 https://www.anthropic.com/news/enterprise-frontier-safeguards 2026-09-01T00:00:00+00:00 Developing Enterprise Frontier Safeguards with our customers 2026-09-01T00:00:00+00:00 https://www.anthropic.com/news/improving-alignment-security-efforts 2026-08-31T00:00:00+00:00 Improving our alignment and security practices 2026-08-31T00:00:00+00:00 https://alignment.anthropic.com/2026/taste/ 2026-08-28T00:00:00+00:00 TASTE: Can AI Models Judge AI Safety Research Proposals? 2026-08-28T00:00:00+00:00 https://www.anthropic.com/news/expanding-support-for-scientists 2026-08-27T00:00:00+00:00 Expanding our support for scientists 2026-08-27T00:00:00+00:00 https://www.anthropic.com/news/model-hardware-standard-research-preview 2026-08-27T00:00:00+00:00 Anthropic is opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical devices, to a first group of scientific research labs and advanced manufacturers. 2026-08-27T00:00:00+00:00 https://www.anthropic.com/research/enabling-independent-research 2026-08-26T00:00:00+00:00 Earlier this year, we ran a pilot giving external researchers access to aggregate, real-world Claude usage data. Three research groups designed their own studies for Anthropic Insights, our privacy-preserving analysis tool. In this post, we share high-level results from those studies and what we learned running this pilot. 2026-08-26T00:00:00+00:00 https://www.anthropic.com/news/wellbeing-research-grants 2026-08-25T00:00:00+00:00 Anthropic is launching a $5 million grant program to fund independent research into how AI impacts users’ wellbeing. 2026-08-25T00:00:00+00:00 https://alignment.anthropic.com/2026/lie-detectors/ 2026-08-21T00:00:00+00:00 Fine-Tuned Lie Detectors Failed to Generalize 2026-08-21T00:00:00+00:00 https://transformer-circuits.pub/2026/interference_effectiveness_helpfulness/index.html 2026-08-21T00:00:00+00:00 We identify interference weights in a 1-layer transformer by measuring their effect on model outputs and loss. 2026-08-21T00:00:00+00:00 https://alignment.anthropic.com/2026/chive/ 2026-08-21T00:00:00+00:00 Would This Change Your Answer? Evaluating Explanations of LLM Behavior in the Wild with Counterfactual Experiments 2026-08-21T00:00:00+00:00 https://www.anthropic.com/research/Claude-accelerates-protein-design 2026-08-18T00:00:00+00:00 In this post, we share two results that show how Claude can help life scientists increase the pace of their research. In the first, we tested Claude’s ability to design protein binders from scratch, a key step in creating protein-based drugs that has historically taken a specialist weeks or months per target. In the second example, we evaluated whether Claude can accelerate chemical analysis. Claude Opus 5, a generally available model, was given NMR and LC-MS data (the data that allows chemists to assess the identity and purity of the compounds they work with). 2026-08-18T00:00:00+00:00 https://www.anthropic.com/news/claude-text-watermark 2026-08-14T00:00:00+00:00 Future Claude models will generate text that contains a watermark. This is a way of determining the likelihood that Claude was involved in writing the text, and we, along with several other major AI providers, are implementing this change to comply with the EU AI Act. In this article, we share answers to some of the questions we’ve received about how our chosen watermarking method works, whether it affects Claude’s outputs, and why we’re making this change. 2026-08-14T00:00:00+00:00 https://www.anthropic.com/research/multiagent-systems 2026-08-13T00:00:00+00:00 We ran experiments on swarms of Claude agents and found coordination failures, collusion, and sabotage. Here, we share what they mean for AI safety. 2026-08-13T00:00:00+00:00 https://www.anthropic.com/research/reviewing-the-evidence-on-worker-retraining-programs 2026-08-12T00:00:00+00:00 An evidence review from Anthropic's Economic Research team 2026-08-12T00:00:00+00:00 https://alignment.anthropic.com/2026/conceptual-reasoning-index/ 2026-08-12T00:00:00+00:00 Introducing the Conceptual Reasoning Index 2026-08-12T00:00:00+00:00 https://www.anthropic.com/research/riemann-zeta 2026-08-10T00:00:00+00:00 An unreleased version of Claude has made strides on a problem related to the Riemann hypothesis. It improved the lower bound for the fraction of zeros of the Riemann zeta function that satisfy the hypothesis, increasing it from 41.6% to 67.2%. 2026-08-10T00:00:00+00:00 https://www.anthropic.com/news/improving-fable-5-s-biology-safeguards 2026-08-07T00:00:00+00:00 We’re making updates to Claude Fable 5’s biology safeguards in a way that substantially reduces fallbacks. 2026-08-07T00:00:00+00:00 https://www.anthropic.com/news/tino-cuellar 2026-08-04T00:00:00+00:00 Tino Cuellar joins Anthropic as Chief Global Affairs Officer 2026-08-04T00:00:00+00:00 https://alignment.anthropic.com/2026/reward-seeker/ 2026-08-01T00:00:00+00:00 Training a Misaligned Reward Seeker 2026-08-01T00:00:00+00:00 https://alignment.anthropic.com/2026/automated-alignment-researchers/ 2026-08-01T00:00:00+00:00 Automated Researchers Can Reliably Mitigate Alignment Failures 2026-08-01T00:00:00+00:00 https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals 2026-07-30T00:00:00+00:00 In a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Below we describe what happened, how it happened, and what we’re changing. We encourage other AI labs to perform similar reviews. 2026-07-30T00:00:00+00:00 https://www.anthropic.com/research/discovering-cryptographic-weaknesses 2026-07-28T00:00:00+00:00 Anthropic researchers find weaknesses in cryptographic algorithms with Claude Mythos Preview 2026-07-28T00:00:00+00:00 https://www.anthropic.com/news/position-open-weights-models 2026-07-27T00:00:00+00:00 Anthropic CEO Dario Amodei on open-weights models 2026-07-27T00:00:00+00:00 https://www.anthropic.com/news/cognizant-anthropic 2026-07-27T00:00:00+00:00 Cognizant embeds Claude across its platforms, with 30,000+ associates trained, and becomes a Global Premier Partner in the Claude Partner Network. 2026-07-27T00:00:00+00:00 https://www.anthropic.com/news/claude-opus-5 2026-07-24T00:00:00+00:00 Opus 5 is a step change improvement for the Opus tier powering long-running agents while delivering improvements in coding and professional work. 2026-07-24T00:00:00+00:00 https://www.anthropic.com/research/project-pilot 2026-07-24T00:00:00+00:00 We worked with Andon Labs on Drone-Bench, a new benchmark testing whether AI models can autonomously fly a drone to locate and follow a person. 2026-07-24T00:00:00+00:00 https://www.anthropic.com/news/anthropic-economic-index-connector 2026-07-22T00:00:00+00:00 We're launching the Anthropic Economic Index connector for Claude, which lets anyone explore the data directly. 2026-07-22T00:00:00+00:00 https://www.anthropic.com/news/economic-futures-research-fund-agenda 2026-07-22T00:00:00+00:00 We’re committing $200 million to the Anthropic Economic Futures Research Fund to support ambitious external research. 2026-07-22T00:00:00+00:00 https://www.anthropic.com/news/donation-public-first-action 2026-07-21T00:00:00+00:00 Anthropic is contributing an additional $20 million to Public First Action, bringing our total support to $40 million. 2026-07-21T00:00:00+00:00 https://www.anthropic.com/news/rare-disease-research-grants 2026-07-20T00:00:00+00:00 Anthropic is sharing a focused call for AI for Science applications centered specifically on rare genetic diseases. Accepted applicants will receive up to $50,000 in Claude credits over six months, with the goal of building a community of researchers looking into how AI can reshape our understanding of rare disease. 2026-07-20T00:00:00+00:00 https://www.anthropic.com/news/claude-for-teachers 2026-07-14T00:00:00+00:00 Introducing Claude for Teachers 2026-07-14T00:00:00+00:00 https://www.anthropic.com/research/how-canada-uses-claude 2026-07-14T00:00:00+00:00 How Canada uses Claude 2026-07-14T00:00:00+00:00 https://www.anthropic.com/news/canadian-ai-research 2026-07-14T00:00:00+00:00 Anthropic is committing $10M to Canadian research institutions to fund the next generation of AI research. 2026-07-14T00:00:00+00:00 https://www.anthropic.com/research/claude-values-models-languages 2026-07-13T00:00:00+00:00 We analyzed 300,000 real conversations to measure the values Claude expresses across models and languages, compressed into four interpretable axes. 2026-07-13T00:00:00+00:00 https://www.anthropic.com/research/claude-plays-robotics 2026-07-09T00:00:00+00:00 Do language models’ strengths transfer to robotics? Can a model perceive a scene, understand a particular robot’s state, and issue actions that reliably effect change in the physical world? We ran tests to find out. 2026-07-09T00:00:00+00:00 https://www.anthropic.com/news/ust-claude 2026-07-09T00:00:00+00:00 UST is bringing Claude to physical AI 2026-07-09T00:00:00+00:00 https://www.anthropic.com/news/ben-bernanke 2026-07-09T00:00:00+00:00 Ben Bernanke appointed to Anthropic’s Long-Term Benefit Trust 2026-07-09T00:00:00+00:00 https://www.anthropic.com/news/hard-questions 2026-07-09T00:00:00+00:00 We're asking the public for their hardest questions about AI, and committing to show our work as we address them. 2026-07-09T00:00:00+00:00 https://www.anthropic.com/news/reflect-with-claude 2026-07-09T00:00:00+00:00 Introducing a new way to reflect on and refine how you use Claude. It lets you easily track and visualize how you use Claude, and decide whether that time aligns with your goals. 2026-07-09T00:00:00+00:00 https://alignment.anthropic.com/2026/modular-pretraining/ 2026-07-08T00:00:00+00:00 Modular Pretraining Enables Access Control 2026-07-08T00:00:00+00:00 https://www.anthropic.com/research/off-switch-dual-use 2026-07-08T00:00:00+00:00 New results on a method of controlling access to potentially dangerous AI capabilities 2026-07-08T00:00:00+00:00 https://transformer-circuits.pub/2026/workspace/index.html 2026-07-06T00:00:00+00:00 We find that Claude maintains a small, privileged set of representations it can report on, control, and reason with, atop a much larger volume of automatic processing. 2026-07-06T00:00:00+00:00 https://www.anthropic.com/research/global-workspace 2026-07-06T00:00:00+00:00 Interpretability research on Claude's internal thoughts. 2026-07-06T00:00:00+00:00 https://www.anthropic.com/news/alberta-government-claude-cybersecurity 2026-07-06T00:00:00+00:00 The Government of Alberta has been using Claude Code with both Opus and Sonnet models to review its systems, find vulnerabilities, and fix them. 2026-07-06T00:00:00+00:00 https://www.anthropic.com/news/fable-safeguards-jailbreak-framework 2026-07-02T00:00:00+00:00 What is and isn't blocked by our cyber classifiers, and a first draft of our jailbreak severity framework 2026-07-02T00:00:00+00:00 https://transformer-circuits.pub/2026/june-update/index.html 2026-06-30T00:00:00+00:00 A short update on turn-averaged sparse autoencoders. 2026-06-30T00:00:00+00:00 https://www.anthropic.com/news/redeploying-fable-5 2026-06-30T00:00:00+00:00 Anthropic is redeploying Claude Fable 5 starting July 1 following the lifting of export controls, with updated cybersecurity safeguards and a new industry jailbreak framework. 2026-06-30T00:00:00+00:00 https://www.anthropic.com/news/claude-science-ai-workbench 2026-06-30T00:00:00+00:00 Claude Science is a customizable app that integrates the tools and packages researchers most often use, produces auditable artifacts, and provides flexible access to computing resources. 2026-06-30T00:00:00+00:00 https://www.anthropic.com/news/claude-sonnet-5 2026-06-30T00:00:00+00:00 Our most agentic Sonnet yet, with top-tier intelligence for coding and everyday professional work. 2026-06-30T00:00:00+00:00 https://www.anthropic.com/research/economic-index-june-2026-report 2026-06-26T00:00:00+00:00 In the latest Anthropic Economic Index report, we look at when people come to Claude, what they produce with it, and how they perceive AI’s impact on their work. 2026-06-26T00:00:00+00:00 https://alignment.anthropic.com/2026/diffuse-ai-control/ 2026-06-23T00:00:00+00:00 Diffuse AI Control on Fuzzy Tasks 2026-06-23T00:00:00+00:00 https://www.anthropic.com/news/introducing-claude-tag 2026-06-23T00:00:00+00:00 Introducing Claude Tag 2026-06-23T00:00:00+00:00 https://www.anthropic.com/research/project-fetch-phase-two 2026-06-18T00:00:00+00:00 We report results from our latest test of whether Claude can help Anthropic employees perform sophisticated robotics tasks. We found that Claude Opus 4.7, operating without human assistance, was about 20 times faster than the fastest human team at all tasks completed by participants less than a year ago. 2026-06-18T00:00:00+00:00 https://www.anthropic.com/news/seoul-office-partnerships-korean-ai-ecosystem 2026-06-17T00:00:00+00:00 Anthropic opens Seoul and announces new partnerships across the Korean AI ecosystem—with the enterprises, startups, and researchers behind some of the most ambitious deployments of Claude. 2026-06-17T00:00:00+00:00 https://www.anthropic.com/research/claude-code-expertise 2026-06-17T00:00:00+00:00 Agentic coding and persistent returns to expertise 2026-06-17T00:00:00+00:00 https://www.anthropic.com/news/fable-mythos-access 2026-06-12T00:00:00+00:00 The US government has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States. 2026-06-12T00:00:00+00:00 https://www.anthropic.com/news/tcs-anthropic-partnership 2026-06-12T00:00:00+00:00 We’re announcing a partnership with Tata Consultancy Services (TCS). TCS will provide Claude to 50,000 of its own employees across 56 countries; build Claude-powered products for clients in financial services, healthcare, the public sector, and other regulated industries; and join the Claude Partner Network. 2026-06-12T00:00:00+00:00 https://www.anthropic.com/news/anthropic-public-record 2026-06-12T00:00:00+00:00 Anthropic Public Record is a national survey of attitudes and opinions towards AI. 2026-06-12T00:00:00+00:00 https://www.anthropic.com/news/dxc-anthropic-alliance 2026-06-11T00:00:00+00:00 We’re announcing a multi-year global alliance with DXC Technology, one of the world’s largest IT services companies. 2026-06-11T00:00:00+00:00 https://www.anthropic.com/news/claude-corps 2026-06-11T00:00:00+00:00 We’re launching Claude Corps, a national fellowship program for people early in their careers who are passionate about extending the benefits of AI to communities across America. 2026-06-11T00:00:00+00:00 https://www.anthropic.com/news/claude-fable-5-mythos-5 2026-06-09T00:00:00+00:00 Today we’re launching Claude Fable 5: a Mythos-class model that we’ve made safe for general use. 2026-06-09T00:00:00+00:00 https://alignment.anthropic.com/2026/agentic-misalignment-summer-2026/ 2026-06-08T00:00:00+00:00 Agentic Misalignment in Summer 2026 2026-06-08T00:00:00+00:00 https://www.anthropic.com/research/n-days 2026-06-08T00:00:00+00:00 N days 2026-06-08T00:00:00+00:00 https://red.anthropic.com/2026/n-days/ 2026-06-08T00:00:00+00:00 Measuring LLMs’ impact on N-day exploits 2026-06-08T00:00:00+00:00 https://www.anthropic.com/research/agents-in-biology 2026-06-08T00:00:00+00:00 Paving the way for agents in biology 2026-06-08T00:00:00+00:00 https://www.anthropic.com/research/making-claude-a-chemist 2026-06-05T00:00:00+00:00 Making Claude a chemist 2026-06-05T00:00:00+00:00 https://www.anthropic.com/research/attack-navigator 2026-06-03T00:00:00+00:00 We’ve spent the past year investigating how threat actors are weaponizing AI to conduct cyber operations. Today, we’re sharing a new analysis that maps these real-world attacks onto the MITRE ATT&CK framework, a database of tactics and techniques used by cyberattackers. 2026-06-03T00:00:00+00:00 https://red.anthropic.com/2026/attack-navigator/ 2026-06-03T00:00:00+00:00 Mapping AI-enabled cyber threats: Insights from the LLM ATT&CK Navigator 2026-06-03T00:00:00+00:00 https://www.anthropic.com/news/services-track-partner-hub 2026-06-03T00:00:00+00:00 In March, we launched the Claude Partner Network, a program for the firms that help enterprises put Claude into production.. Today, we’re announcing two new components that make this ecosystem easier for customers to navigate. 2026-06-03T00:00:00+00:00 https://www.anthropic.com/news/AI-enabled-cyber-threats-mitre-attack 2026-06-03T00:00:00+00:00 As AI transforms the nature of and methods behind cyberattacks, how well do the techniques and frameworks used by the security community hold up? In a new report, we seek to answer that question. 2026-06-03T00:00:00+00:00 https://www.anthropic.com/news/expanding-project-glasswing 2026-06-02T00:00:00+00:00 We’re extending Project Glasswing to approximately 150 new organizations in more than fifteen countries 2026-06-02T00:00:00+00:00 https://transformer-circuits.pub/2026/may-update/index.html 2026-06-01T00:00:00+00:00 A short update on understanding features through downstream connections. 2026-06-01T00:00:00+00:00 https://www.anthropic.com/news/confidential-draft-s1-sec 2026-06-01T00:00:00+00:00 Anthropic has confidentially submitted a draft S-1 registration statement to the Securities and Exchange Commission 2026-06-01T00:00:00+00:00 https://www.anthropic.com/news/series-h 2026-05-28T00:00:00+00:00 Anthropic has raised $65 billion in Series H funding led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. 2026-05-28T00:00:00+00:00 https://www.anthropic.com/news/claude-opus-4-8 2026-05-28T00:00:00+00:00 Our latest model, Claude Opus 4.8, is 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 work. 2026-05-28T00:00:00+00:00 https://www.anthropic.com/research/coding-agents-social-sciences 2026-05-27T00:00:00+00:00 Results from a survey of 1,260 social scientists about AI and coding agent use. 2026-05-27T00:00:00+00:00 https://www.anthropic.com/news/milan-office-opening 2026-05-27T00:00:00+00:00 We're opening a new office in Milan, our sixth in Europe. 2026-05-27T00:00:00+00:00 https://www.anthropic.com/news/kiyoung-choi-representative-director-anthropic-korea 2026-05-26T00:00:00+00:00 KiYoung Choi is joining Anthropic as Representative Director of Korea, ahead of the opening of our Seoul office. 2026-05-26T00:00:00+00:00 https://www.anthropic.com/news/chris-olah-pope-leo-encyclical 2026-05-25T00:00:00+00:00 The full text of Chris Olah's remarks on the Pope's encyclical on AI 2026-05-25T00:00:00+00:00 https://red.anthropic.com/2026/cvd/ 2026-05-22T00:00:00+00:00 Anthropic's coordinated vulnerability disclosure dashboard 2026-05-22T00:00:00+00:00 https://red.anthropic.com/2026/exploit-evals/ 2026-05-22T00:00:00+00:00 Measuring LLMs’ ability to develop exploits 2026-05-22T00:00:00+00:00 https://www.anthropic.com/research/glasswing-initial-update 2026-05-22T00:00:00+00:00 An early update on what we've learned from Project Glasswing. 2026-05-22T00:00:00+00:00 https://alignment.anthropic.com/2026/sleight-bench/ 2026-05-19T00:00:00+00:00 SLEIGHT-Bench: Finding Blind Spots in AI Monitors 2026-05-19T00:00:00+00:00 https://www.anthropic.com/news/anthropic-kpmg 2026-05-19T00:00:00+00:00 KPMG integrates Claude across its core business and workforce of more than 276,000 in strategic alliance 2026-05-19T00:00:00+00:00 https://www.anthropic.com/news/widening-conversation-ai 2026-05-19T00:00:00+00:00 Widening the conversation on frontier AI 2026-05-19T00:00:00+00:00 https://www.anthropic.com/news/anthropic-acquires-stainless 2026-05-18T00:00:00+00:00 Anthropic acquires Stainless 2026-05-18T00:00:00+00:00 https://www.anthropic.com/research/2028-ai-leadership 2026-05-14T00:00:00+00:00 Our views on the AI competition between the US and China. 2026-05-14T00:00:00+00:00 https://www.anthropic.com/research/teaching-claude-why 2026-05-08T00:00:00+00:00 New research on how we've reduced agentic misalignment 2026-05-08T00:00:00+00:00 https://transformer-circuits.pub/2026/nla/index.html 2026-05-07T00:00:00+00:00 We train Claude to translate its internal state into natural language. 2026-05-07T00:00:00+00:00 https://www.anthropic.com/research/anthropic-institute-agenda 2026-05-07T00:00:00+00:00 At The Anthropic Institute (TAI), we’ll be using the information we can access from within a frontier lab to investigate AI’s impact on the world, and sharing our learnings with the public. Here, we’re sharing the questions that drive our research agenda. 2026-05-07T00:00:00+00:00 https://www.anthropic.com/research/donating-open-source-petri 2026-05-07T00:00:00+00:00 Updating Petri to version 3.0 and donating it to Meridian Labs 2026-05-07T00:00:00+00:00 https://www.anthropic.com/research/natural-language-autoencoders 2026-05-07T00:00:00+00:00 Turning Claude's thoughts into text 2026-05-07T00:00:00+00:00 https://alignment.anthropic.com/2026/msm/ 2026-05-05T00:00:00+00:00 Model Spec Midtraining: Improving How Alignment Training Generalizes 2026-05-05T00:00:00+00:00 https://transformer-circuits.pub/2026/headvis/index.html 2026-05-04T00:00:00+00:00 We develop an interactive visualization tool to help us understand the behaviors of attention heads in language models. 2026-05-04T00:00:00+00:00