French AI lab Mistral AI has released Mistral Large 4 (ML4), a new large multimodal model intended to compete with both American and Chinese rivals, in line with what French President Emmanuel Macron has described as a third way in AI.
Nicknamed Le Chonk in reference to its 1 trillion parameters, ML4 is positioned as an alternative to closed models that can be unplugged and to open models that are often made in China. For now, it is accessible only through a public guardrail endpoint, with Mistral planning to release the weights in three weeks once safety testing is complete.
Mistral VP Science Pierre Stock told TechCrunch that the company will work with trusted partners and governments in the meantime to ensure the open weights can be used for defense rather than malicious attacks. He added that open-weight models are easier to audit, a growing priority for Mistral’s enterprise and institutional customers.
Stock said ML4 was trained entirely on Mistral’s own compute using 4,000 Nvidia GPUs, two to three times fewer than Chinese competitors and significantly fewer than closed-source rivals. With benchmark results pending, Mistral hopes the model will lead its open-weight class and outperform closed models in targeted areas.
According to Stock, optimized use cases include cybersecurity, finance and chip design, the latter core to backers ASML, which led Mistral’s Series C, and Samsung, which led its Series D last month at a €21 billion valuation.
