Mistral released a brand-new AI model on Tuesday, which the company says ranks among the world's top open-source models. The French firm introduced Mistral Large 4 (abbreviated as ML4, codenamed le Chonk) as a large model with 1 trillion parameters, delivering particularly strong performance in cybersecurity, code writing, manufacturing, finance, and multimodal-related tasks.

Mistral has long been regarded as the only European company capable of rivaling OpenAI and Anthropic. The company completed a €3 billion (equivalent to US$3.4 billion) Series D funding round in September, reaching a valuation of €21 billion. This funding round solidified its position as one of Europe's best-funded and highest-valued AI startups.

Mistral stated that Mistral Large 4 is now available in preview for developers, cybersecurity leaders, and government agencies, with a full broad release scheduled for later this month. The ML4 model was trained over two months using 4,000 Nvidia Grace Blackwell GPUs within Mistral's own data center in Europe.

In a statement, Mistral said that once the core parameters are officially released, ML4's comprehensive benchmark performance will rank among the world's top open-weight models. However, this new model still lags behind the industry frontier in areas such as code writing.

Guillaume Lample, Mistral's co-founder and chief scientist, said: "After completing the Series D funding round, as the scale of training compute continues to increase, model performance will be further enhanced." He added: "In addition, the model's cyber defense capabilities can help enterprises and governments defend against malicious attackers—attackers who bypass the safety guardrails of closed-source models and exploit them to carry out cyberattacks."