Mistral AI Updates: May 15, 2026

1. Mistral developing cybersecurity AI model for European banks locked out of Anthropic’s Mythos

Mistral. Bloomberg reported that Mistral is in active talks with European banks about a homegrown answer to Anthropic’s Mythos, the highly restricted model known for surfacing cyber vulnerabilities at scale, currently available only to a handful of select banks, security firms, and tech companies. Mistral had already worked with banking customers on AI-based vulnerability discovery before Mythos launched and is now productizing the work for broader European distribution. Anthropic’s gating of Mythos has triggered a wider race (including OpenAI) to ship comparable defensive-security models, with European financial-sector buyers under pressure to find a Mythos alternative. Source