The Mistral 1T model has arrived, and Europe is still in the mix. On Tuesday, French AI lab Mistral AI released Mistral Large 4 (ML4), a new large multimodal model aiming to leapfrog both American and Chinese rivals — following what French president Macron described as “a third way in AI.”
Amid a growing divide between closed models that can be unplugged and open models that are often made in China, Mistral is positioning ML4 as an alternative to both. Nicknamed Le Chonk in reference to its 1 trillion parameters, the Mistral 1T model is definitely not small — but it is not an open-weight model yet.
What Makes the Mistral 1T Model Different
For the time being, ML4 can only be accessed via a public guardrail endpoint. However, Mistral plans to make its weights available in just three weeks, after safety testing is complete.
“In the meantime, we’ll work with trusted partners and governments to make sure that the open source weights can be used to defend, but not to [perform] malicious attacks,” Mistral VP Science Pierre Stock said.
Security concerns have been mounting in recent months, particularly among Mistral’s core audience — enterprises and institutions. At the same time, an open-weight model is easier to audit, Stock said.
That balance between security and transparency sits at the heart of the Mistral 1T model’s design. Mistral wants organizations to be able to inspect and trust the system without handing bad actors an easy weapon.
Trained on Mistral’s Own Compute
Another important behind-the-scenes aspect is that the Mistral 1T model was trained entirely on Mistral’s compute, using only 4,000 Nvidia GPUs.
That is “two to three times less than our Chinese competitors, and significantly less than the closed source competitors,” Stock said.
The efficiency claim matters for several reasons. Fewer GPUs means lower training costs, reduced energy demands, and a smaller hardware footprint than many rival systems. For enterprises weighing the total cost of adopting a frontier model, those savings can influence long-term decisions.
Benchmarks and Best-in-Class Ambitions
With benchmark results still pending, Mistral hopes ML4 will be best in class among open-weight models, especially outside of China, but not only, Stock said.
Thanks to focused training, it could also outperform closed models in specific areas that are key to its customers, and where multimodal capabilities can add value.
According to Stock, ML4’s optimized use cases include cybersecurity and finance, but also chip design. That last area is core to two of Mistral’s main backers — Dutch giant ASML, which led its Series C, and Samsung, which led its Series D last month at a €21 billion valuation (about $24.39 billion).
A Frontier Lab, Not Just an Inference Provider
At the time of that funding round, the company tried to convey that its decision to host Chinese models wasn’t a pivot into becoming a mere inference provider. With Le Chonk in its corner, the company believes it should still be considered a frontier lab as well.
That distinction matters to Mistral’s identity. A frontier lab builds models at the leading edge of capability, while an inference provider mainly runs models built by others. Mistral wants to be seen as the former, and the Mistral 1T model is its strongest argument yet.
Why the Mistral 1T Model Matters
The Mistral 1T model arrives at a moment when enterprises and governments are weighing the trade-offs between open and closed AI systems. Closed models offer tighter control but can be unplugged at any time. Open models offer auditability but often come with geopolitical baggage.
Mistral is betting that ML4 can occupy the middle ground: a European-built, multimodal frontier model that is open enough to audit but guarded enough to satisfy security-conscious customers.
The three-week timeline for releasing the weights will be a key test. If Mistral delivers, the Mistral 1T model could become a credible alternative for organizations that want frontier performance without depending entirely on closed providers or Chinese open-weight releases.
Benchmark results are still pending, and those numbers will shape how the market receives ML4. Until then, Mistral’s claims rest on its training efficiency, its targeted use cases, and its promise to open the weights soon.
For now, the Mistral 1T model stands as Europe’s most ambitious answer to the open-versus-closed AI divide. Whether it leapfrogs rivals or simply holds its ground, it keeps Europe in a race that many assumed was already decided.
