Mistral AI Releases 1-Trillion-Parameter Model Mistral Large 4, With Open Weights Coming in Three Weeks

French AI lab Mistral AI released Mistral Large 4 (ML4) on Tuesday, October 6, 2026 — a large multimodal model with one trillion parameters that the company says is designed to compete with both American and Chinese AI rivals.

Nicknamed “Le Chonk” for its scale, ML4 is not yet available as an open-weight model. For now, it can only be accessed through a public guardrail endpoint. Mistral plans to release the model weights within three weeks, pending the completion of safety testing.

“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 of Science Pierre Stock told TechCrunch.

ML4 was trained entirely on Mistral’s own compute infrastructure using 4,000 Nvidia GPUs — a figure Stock described as “two to three times less than our Chinese competitors, and significantly less than the closed source competitors.” Benchmark results are still pending, but Mistral expects ML4 to rank among the top open-weight models globally, and potentially outperform closed models in specific use cases where multimodal capabilities add value.

Mistral has identified cybersecurity, finance, and chip design as optimized use cases for ML4. The chip design focus aligns with the interests of two of the company’s major backers: Dutch semiconductor equipment maker ASML, which led Mistral’s Series C, and Samsung, which led its Series D last month at a €21 billion valuation (approximately $24.39 billion).

The release comes as Mistral seeks to reinforce its standing as a frontier AI lab — not merely an inference provider — following scrutiny over its decision to host Chinese models. ML4 also reflects what French President Emmanuel Macron has described as “a third way in AI,” positioning European-developed open-weight models as an alternative to both closed proprietary systems and Chinese open-source offerings.