French security firm Orange Cyberdefense plans to use Mistral AI’s latest model to test clients’ systems for weaknesses, investigate breaches and speed up its response to cyberattacks. The company joined the design partner program for Mistral Large 4 on Thursday, becoming the first European managed security services provider to take part.
What will Orange Cyberdefense use Mistral Large 4 for?
The Orange unit will introduce the model gradually across its services. Its first uses will include technical audits and penetration testing, in which security specialists try to break into a client’s systems to find vulnerabilities.
The model will also help teams assess how attackers might gain access to those systems. When a breach occurs, it will help investigators work out what happened. The aim is to help security staff understand incidents and respond faster.
Mistral launched Large 4 in early October. The model has one trillion parameters. The company has since made its weights available for download, so developers can run and adapt the model themselves.
Who can use the cybersecurity version of Mistral Large 4?
A cybersecurity-focused version is now being tested by a small group of partners. Those partners may use it only on systems they own or have permission to assess.
Where will Orange Cyberdefense run its security models?
Orange Cyberdefense will run its security models on computers in Mistral’s European data centers. The company says this arrangement will help clients retain control over their data, which is becoming a bigger concern as firms use AI for sensitive security work. Mistral has made data control and deployment flexibility part of its argument for open models over relying entirely on closed AI systems.
“This alliance between Orange Cyberdefense and Mistral AI strengthens the foundation of cybersecurity augmented by trustworthy AI,” said Anis Trabelsi, head of innovation and AI at Orange Cyberdefense.
The partnership builds on the firm’s existing use of AI. Since April 2025, it has used a tool from Qevlar AI in its security operations centers, where staff monitor clients’ systems for threats.
The companies did not disclose the financial terms of the agreement. The rollout will initially target business and public-sector clients.
This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.