Full Summary
This Tuesday morning, Anthropic officially launches Claude Fable 5, its most powerful public AI model to date. Both Anthropic's own announcement and reports from 9to5Mac confirm this "Mythos-class" model is now available to customers. Claude Fable 5 is designed for general use, excelling in software engineering, knowledge work, vision, and scientific research. It scored 80.3% on the SWE-Bench Pro software engineering benchmark and was the first AI model to cross the 90% threshold on Hex's analytical benchmark, outperforming Claude Opus 4.8 by over 10%. Moneycontrol.com highlights its prowess in complex research and long-horizon problem-solving. However, a key development, confirmed by The Independent and Crypto Briefing, is the inclusion of significant safeguards. For high-risk requests related to cybersecurity, biology, chemistry, or model distillation, Fable 5 will defer to Claude Opus 4.8. These safeguards are expected to trigger in less than 5% of sessions. Alongside Fable 5, Anthropic also launched Claude Mythos 5, a version with fewer safeguards, accessible only to a select group of cyberdefenders and trusted cybersecurity partners, including the US government through Project Glasswing. Mint and outlookbusiness.com previously detailed how the original Mythos model, initially withheld due to its unprecedented capabilities, had already discovered thousands of high-severity vulnerabilities in major software, including a 27-year-old flaw in OpenBSD. In related news, Apple has introduced Siri AI, a more conversational and context-aware assistant powered by Apple Intelligence. Let's Data Science notes its integration across iPhone, iPad, Mac, Apple Watch, and Vision Pro. However, Tech Times reports that Siri AI will not be available to approximately 450 million EU users due to regulatory issues with the Digital Markets Act. Finally, OpenAI CEO Sam Altman is urging U.S. lawmakers to avoid requiring federal approval before AI models are released, advocating instead for increased funding for AI testing at the Department of Commerce, as reported by Let's Data Science. These AI advancements mean your next software update could be more secure, but also that regulatory battles could limit access to cutting-edge AI features depending on where you live.