Overview
- BNP Paribas and Mistral announced an expanded phase of their tie-up on Tuesday, May 26, 2026, shifting the partnership’s priority to defending banks from advanced AI-driven cyber threats.
- Mistral is developing a dedicated cybersecurity model pitched at European banks, but its commercial launch date and broad availability have not been confirmed by the company.
- The pair have worked together since September 2023 and formalised a multi-year agreement in July 2024, and Mistral engineers are already embedded in BNP teams to co-develop tools used for assistants, document extraction, and compliance.
- The move was accelerated by reports that Anthropic’s Mythos model achieved high exploit success in tests and identified a long-standing OpenBSD flaw, which bankers say shows AI can find vulnerabilities far faster than human teams.
- Bank officials and European regulators are planning new processes to triage and patch large volumes of AI-found flaws in parallel, and wider adoption of a regional model could create shared threat intelligence that strengthens collective defenses.