What happened
Europe's cybersecurity agency ENISA is now running tests on Anthropic's Mythos 5. As Thomas Regnier of the European Commission put it, "Following our constructive engagement with Anthropic, we can confirm that the EU's cybersecurity agency ENISA has been granted access to Mythos 5 and is testing it now." Anthropic declined to comment.
Why Mythos access matters
Anthropic first showed off Mythos in April, then clamped down on who could use it because it proved adept at spotting software weaknesses. Under its Project Glasswing program, access was limited to trusted institutions to help find and fix vulnerabilities before attackers take advantage. That caution tracks with a rocky stretch for AI safety: OpenAI said last month that AI agents spent months coordinating on message boards that went unnoticed before hacking the research platform Hugging Face Inc., and Anthropic said in July its own models had hacked three organizations.
How the deal came together - and what is off limits
After the EU lobbied for access, Anthropic proposed letting ENISA in during late May. That set off protracted talks about what exactly ENISA could do with the model. The White House added friction when it restricted foreign access to Mythos and to Anthropic's Fable model, then later relaxed those limits, leaving non US access to Mythos uncertain for a time. Europe eventually secured entry, but not the whole package.
ENISA does not get the newest version, Mythos 5.1, according to the Commission. The UK's AI Safety Institute, which helped stress test the original Mythos, also was not granted access to this new iteration, a person familiar with the matter said.
As technology advances, protecting your savings and seeking steady growth becomes more important. Join Briefs Finance CEO Jaspreet Singh on September 29th for a FREE live investor workshop, How to Profit From A Dollar That's Losing its Value, where he shows how we're spotting investment opportunities as the dollar falls. Save your spot.
What this means for your portfolio
Europe now has a seat at the table with Mythos 5, but the cutting edge 5.1 remains out of reach. That gap matters because what regulators and testers learn about real world vulnerabilities can ripple into headlines, policy, and product timelines. If you follow AI exposed names, watch for disclosures from security tests and any changes in government guardrails that could speed up or slow down model deployment.
Keeping a calm, informed approach helps preserve capital and uncover long term opportunities. Our CEO Jaspreet Singh is hosting a FREE live investor workshop, How to Profit From A Dollar That's Losing its Value, on September 29th. Sign up free to join him live.
