Mistral unveils Le Chonk as alternative to top AI models from China, US
Also known as Mistral Large 4.
Mistral AI advertisement in New York. Photo by REUTERS Shannon Stapleton
Follow us
- Mistral unveiled Le Chonk, officially Mistral Large 4, after about five months without a major model release.
- The company says the open-weight model outperforms some rivals, including in cybersecurity, but did not specify benchmarks.
- Cybersecurity experts and government authorities will test a less restricted version before public release on October 27th.
- Mistral says Large 4 can support coding, finance, geospatial analysis, manufacturing, product design, and chip design.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
French AI startup Mistral has unveiled a new AI model, ending a gap of about five months between major releases, and said it outperforms many rival open-weight models.
The three-year-old company, widely viewed as Europe's strongest AI challenger to US and Chinese competitors, has not released a new model since May. In August, it began offering third-party models to customers, starting with Beijing-based Z.ai's GLM 5.3.
"The model we're actually announcing today is actually above the Chinese models on certain aspects, including cyber. So the narrative that Europe cannot compete is something that is not true," CEO Artur Mensch said at a launch event in Abu Dhabi, without specifying which Chinese models or benchmarks he was referring to.
Mistral said Le Chonk, officially named Mistral Large 4, will be made publicly available on October 27th.
Before the release, cybersecurity experts and government authorities will be given access to a version with fewer safety restrictions to test its capabilities, the company said.
Pierre Stock, Mistral's vice president of science, told Reuters that the model ranks among the strongest open-weight, systems available.
Open-weight means users can download the model from platforms and run it on their own hardware. Many Chinese models are open-weight, while top US models tend to be closed and accessible only through their developers' infrastructure.
Safety risks contained
Stock told Reuters the model had tried to go beyond its testing environment, but that this was expected and the company was able to prevent it.
OpenAI and Anthropic, whose most advanced models have demonstrated similar behaviour in testing, have restricted access to their most cyber-capable systems.
After executives at US AI companies warned last month that AI could pose an existential threat to humanity, Mensch dismissed such arguments as self-serving.
Mistral, which raised €3 billion ($3.4 billion) in its latest funding round, has positioned itself as a safer, European alternative.
“Closing the gap”
Mistral said Large 4 is "closing the gap with frontier models" in areas including coding, finance, geospatial analysis, manufacturing and product design.
The company trained the model on its own infrastructure.
Stock said Mistral's goal was to partner "very deeply" with backers including ASML and new-investor Samsung "across the entire value chain".
He said Large 4 is capable of designing computer chips, but declined to comment on whether Mistral is currently working with Samsung on chip development.