‹ Back

Mistral Targets European Banks With Mythos Rival

 

French AI startup Mistral is developing a cybersecurity tool for European banks locked out of Anthropic's restricted Mythos model.

 

 

Mistral in Talks With European Banks on AI Cybersecurity Tool

 

French artificial intelligence startup Mistral AI is in discussions with European banks about deploying its own answer to Anthropic's Mythos — the limited-access AI model capable of uncovering cybersecurity vulnerabilities at unprecedented speed and scale. Bloomberg News first reported the development on Wednesday, May 13, citing people familiar with the matter who asked not to be identified as the information is private.

 

Mistral had already been working with its banking clients on using AI to identify security flaws before Mythos was released, but is now developing an off-the-shelf iteration of a product it can roll out more widely, according to one of the sources cited by Bloomberg. The timeline for release of the model has not been confirmed, but the company has held active discussions about it with the European banking sector.

 

Europe's banks, lacking access to Mythos, are under pressure to detect and fix vulnerabilities that could be exploited by AI tools. Approximately 40 tech firms and institutions have initial access to Mythos to bolster their systems, but these do not include most central banks and governments.

 

Mythos Access Gap Leaves Financial Sector Exposed

 

Among the organizations with confirmed access are JPMorgan Chase, named publicly as a launch partner, along with Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley, according to Reuters sources and company executives. 

 

The few large banks with access are uncovering vulnerabilities in their IT systems and sharing information with smaller banks that cannot access the model, Reuters reported. One finding from Mythos is that the model can create a high-risk vulnerability by combining several lower-risk weaknesses. The number of low- to moderate-ranked vulnerabilities found in banks' systems by Mythos ranges from several hundred to thousands.

 

Anthropic CEO Dario Amodei stated on May 5 that financial services companies and other organizations have six to 12 months to fix vulnerabilities in their software before Chinese AI models develop capabilities equal to those of Mythos.

 

Amodei said Mythos had uncovered tens of thousands of vulnerabilities and that if code is rewritten with models like Mythos, "we could be in a better position than we started in because we fixed all these bugs." Nitin Seth, co-founder and CEO of the data and AI services firm Incedo, described the situation bluntly: "This is a wake-up call because cyber risk is moving to machine speed, while much of bank defense still operates at human speed."

 

For institutions without Mythos access, Anthropic has released recommendations for improving defenses and is offering a separate program called Claude Security, which scans for vulnerabilities and is available to a wider range of organizations.

 

OpenAI Enters Cybersecurity Race With Daybreak

 

Mistral's move follows OpenAI's announcement of a competing initiative. On May 11, 2026, OpenAI CEO Sam Altman announced on X: "OpenAI is launching Daybreak, our effort to accelerate cyber defense and continuously secure software. AI is already good and about to get super good at cybersecurity; we'd like to start working with as many companies as possible now to help them continuously secure themselves."

 

Unlike Mythos, Daybreak is publicly available, and companies can request an assessment of their security risks. Cloudflare, Cisco, CrowdStrike, Oracle, and Zscaler are among the companies already using the technology, OpenAI confirmed. Daybreak is designed to operate in three stages: first, prioritizing high-impact threats through AI reasoning; second, generating and testing risks within an enterprise environment with scoped access and monitoring; and third, delivering audit-ready evidence to help enterprises track, validate, and remedy vulnerabilities.

 

At the government level, the European Commission is currently in discussions with OpenAI regarding access to its advanced AI models for identifying cybersecurity vulnerabilities. Commission spokesperson Thomas Regnier confirmed that OpenAI proactively approached the EU and that discussions are underway around potential next steps, including possible access to the company's new model. Discussions with Anthropic are also continuing but have not yet reached the same stage.

 

Google Documents First AI-Developed Zero-Day Exploit

 

The competitive scramble among AI firms over cybersecurity tooling is running in parallel with documented evidence of AI being weaponized by criminal actors. Google's Threat Intelligence Group reported that, for the first time, it has identified a threat actor using a zero-day exploit believed to have been developed with AI. The criminal threat actor planned to use it in a mass exploitation event, but Google's proactive counter-discovery may have prevented its use.

 

The zero-day was designed to bypass two-factor authentication on an open-source web-based system administration tool and was developed by a prominent cybercrime group. The exploit was implemented in a Python script. John Hultquist, chief analyst at Google Threat Intelligence Group, told CyberScoop: "We finally uncovered some evidence this is happening." Hultquist further stated that "there's a misconception that the AI vulnerability race is imminent. The reality is that it's already begun. For every zero-day we can trace back to AI, there are probably many more out there."

 

Google confirmed that neither Gemini nor Anthropic's Mythos was the AI model used by the attackers. AI-enabled entities increased cyberattacks by 89% in 2025 from a year earlier, according to cybersecurity firm CrowdStrike. The convergence of restricted access to defensive AI tools, an accelerating offensive threat landscape, and the entry of multiple AI providers into the cybersecurity market has placed European financial institutions in particular at the center of a rapidly widening capability gap — one that Mistral, OpenAI, and others are now moving to fill.