Mistral on Tuesday unveiled a new AI model it said ranked among the best open models in the world, as Western companies race to close the capability gap with Chinese developers in the market.
Mistral Large 4 (ML4), also known as le Chonk, is a 1-trillion-parameter model that is particularly effective at tasks related to cyber, coding, manufacturing, finance and multimodal, the French company said.
Long considered Europe’s only answer to OpenAI and Anthropic, Mistral raised a 3 billion euros ($3.4 billion) Series D in September at a 21 billion euro valuation. The round cemented the company’s status as one of the best-funded and most valuable AI startups in the region.
Access to so-called open AI models — the most capable of which have been Chinese — has become a key flashpoint in the intensifying race for technological supremacy between the U.S. and China.
Open AI models can be modified and self-hosted, in contrast to leading closed systems produced by OpenAI and Anthropic. Chinese open models are increasingly being adopted around the world.
Mistral Large 4 is being launched in preview to developers and “cybersecurity leaders,” alongside state authorities, before it is released more widely later this month, the company said.
Rise of open models
ML4 was trained on 4,000 Nvidia Grace Blackwell GPUs over two months that were deployed in Mistral’s own data centers in Europe.
When its core parameters are released, ML4 will rank among the top open-weight models globally on aggregate benchmark performance, Mistral said in a statement. The company added that ML4 was the strongest open-weight model developed outside China by a “substantial margin.”
The new model still lags behind the frontier in areas such as coding.
“The model capabilities will further improve as we scale up our training capacity, following our Series D fundraise,” said Guillaume Lample, co-founder and chief scientist of Mistral.
“Further, the cyber defense capabilities will enable enterprises and governments to defend themselves against threat actors that are jailbreaking closed models to perform cyber attacks,” he added.
Open models were thrust into the spotlight earlier this year when it emerged that startup Hugging Face turned to Chinese-based Z.ai’s GLM 5.2 to defend against a breach from rogue OpenAI agents. Hugging Face initially turned to leading U.S. closed systems to defend itself, but was blocked by guardrails that mistook its actions for an attack.
Nvidia-backed American startup Reflection AI unveiled its first open model on Monday, which it claimed matched the capability of GLM 5.2, which was surpassed by GLM 5.3 in August.