“Mistral Large 3 was trained on a wide variety of languages, making advanced AI useful for billions who speak different native languages,” the company said.
The French artificial intelligence (AI) champion Mistral AI has rolled out new, smaller frontier models that operate in multiple languages to challenge its rivals in the United States.
The company, which launched two years ago, released on Tuesday Mistral Large 3, which the startup claims maintains the same level of performance in “a large number of languages,” particularly European ones.
“Most AI labs focus on their native language, but Mistral Large 3 was trained on a wide variety of languages, making advanced AI useful for billions who speak different native languages,” the company said in a press release.
It is now multimodal, meaning it can read documents other than text, such as audio, images, and video. This feature puts it in the same category as Google’s Gemini 3, launched several weeks ago, and according to benchmarks, is considered one of the best AI models at the moment.
The company Mistral Large 3 is engineered for robotics, autonomous drones, and small on-device applications without network access, as well as the world’s largest enterprise agentic workflows.
Ministral
The large model is available in a series of nine “small models,” called “Ministral,” that can be run directly on devices without the need to connect to the internet.
The next wave of AI won’t be defined by sheer scale, but by ubiquity - by models small enough to run on a drone, in a car, in robots, on a phone or a computer laptop. Small models deliver advantages for most real-world applications: lower inference cost, reduced latency, and domain-specific performance,” the company said.
The benefit of smaller AI models is that they require less computing power, resources and less expensive chips than large language models (LLMs). It also means that they can be faster and operate better in certain environments.
Mistral said that this includes robotics, where Mistral AI’s edge solutions work without Wi-Fi, giving operators instant, on-site diagnostics - using live sensor data and facility-specific repair logs to fix issues right on the shop floor. Its technology can also help in emergencies, so that drones can survive without wifi in dead zones.
The company also said its models were open source, meaning they are customizable models for developers, “making frontier AI accessible regardless of your native language,” Mistral said.
