Anthropic seems to understand that making certain topics off-limits for Fable 5 is something of a double-edged sword. The company writes that “the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors.” That puts Anthropic in the somewhat awkward position of having to judge who is and is not trustworthy enough to have access to a model that it says has potentially dangerous capabilities. The company says it will be periodically expanding its existing Project Glasswing program “in consultation wit
Anthropic says these topics are too dangerous to let its Fable 5 model talk aboutWith this expansion of Anthropic’s Project Glasswing initiative, organizations in “several industries that weren’t well represented” in the initial cohort, like power, water, and healthcare, will get access to the model so they can use it to find security vulnerabilities. …
Anthropic expands Glasswing as it promises public Claude Mythos-class model releases Zac Hall | Jun 2 2026 - 7:01 am PT | Jun 2 2026 - 7:01 am PT Anthropic unveiled its Claude Mythos AI model in April, only sharing access with select partners like Apple. …
… They’ve also shaped our thinking about the very purpose of Project Glasswing. Cheap, fast AI models with powerful cyber capabilities are around the corner. We want Project Glasswing to spur institutions toward operating norms that reflect this reality. …
… The Glasswing collaboration model emphasizes shared learning across participants. …
… Anthropic says Fable 5 operates on the “same underlying model” as Mythos 5, which is coming out of its monthslong “Mythos Preview” period today, but only for “a small group of cyberdefenders” judged trustworthy through the existing Project Glasswing . …
… Add TweakTown as a source on Google Similar News Stories NVIDIA deploys GPT-5.5-powered Codex to 10,000 employees, with engineers calling results 'mind-blowing' Anthropic's Claude Mythos found 271 Firefox vulnerabilities - CTO calls it just as capable as 'elite security researchers' AI Agents like … …
… Project Glasswing: An initial update An early update on what we've learned from Project Glasswing.
… What policies and business models meaningfully shift that balance? How do free or open weight models contribute to this dynamic? …
… Daybreak builds on the success of OpenAI's earlier GPT-5.4-Cyber model, which helped remediate over 3,000 security flaws. It uses Codex Security to create an editable threat model from a company's software repository and focuses on the most realistic attack paths. …