The timing is almost too perfect. On May 13, the ECB publicly warned eurozone lenders that Anthropic’s Mythos — the agentic model that auto-discovers zero-days at machine speed — is already being used to probe them. The same day, Bloomberg reported Mistral is in talks with European banks to ship a rival they can actually buy.
What it actually is
Not a chatbot. It’s a closed-deployment offensive-security model — an AI that crawls a bank’s codebase, chains exploits across systems, and surfaces vulnerabilities before attackers do. Mistral has been doing bespoke versions of this with HSBC and BNP Paribas for months. The new move: an off-the-shelf product any European bank can deploy without begging Anthropic for Project Glasswing access, which today is gated to a handful of US tech firms, cyber vendors, and select banks.
Why it matters
This is the first time a European frontier lab has gone head-to-head with a US lab on a national-security-grade capability. ASML, which led Mistral’s €1.3B round last September at a €12B valuation, suddenly reads less like a chip story and more like a sovereignty bet. CEO Arthur Mensch said the quiet part out loud at a French parliamentary inquiry: Europe cannot let bug-finding on sensitive systems sit behind an American allowlist. No release date. No pricing. Just a queue forming at the door.
You Might Also Like
- Cursor Composer 2 Takes on Anthropic and Openai With a 0 50 m Token Coding Model and the Benchmarks Back it up
- Anthropics Next ai Model was Never Supposed to go Public Heres What the Claude Mythos Leak Tells us
- Project Glasswing Anthropic Deploys a Restricted ai Model Across 12 Tech Giants to Hunt Zero day Bugs
- Mistral Workflows Ships Temporal Powered ai Orchestration Already Running at Asml and cma cgm
- Taalas Just Burned an ai Model Into Silicon and the Numbers are Wild
Top AI Product 
Leave a comment