Mistral’s New ‘Le Chonk’ AI Model Is Big, Open and Built for Agents

Mistral AI launched Mistral Large 4, its new flagship open-weight model, in public preview on Tuesday ahead of an October 27 rollout, with chief scientist Guillaume Lample and VP of science Pierre Stock saying it targets agentic, cybersecurity and geospatial tasks. Mistral says the model, nicknamed "le Chonk," reaches state-of-the-art performance on agentic benchmarks including SWE-bench and Terminal Bench, processes both text and images, and was trained on more than 160 languages including every official European Union language. The company says Large 4 ranks among the best open-weight models worldwide and outperforms European and US models while catching up to Chinese competitors that have dominated the open-weight sector.

New AI models are coming, and they don’t stop coming. On the heels of the recent rollout https://www.cnet.com/tech/services-and-software/anthropic-and-openai-drop-new-high-efficiency-models/ of Claude Opus 5.5 and OpenAI’s GPT-6 Sol and GPT-6 Luna, Mistral AI is launching a new flagship model, Mistral Large 4 https://mistral.ai/news/mistral-large-4/ , with a public preview available on Tuesday. A French company known for its open-source and open-weight LLMs, Mistral AI reports that the Large 4 model “ranks among the best open-weight models of the world” on aggregated metrics and also outperforms European and US models, while also catching up to Chinese competitors, who have been dominating the open-weight sector. More from CNET I talked with Mistral Chief Scientist Guillaume Lample and VP of Science Pierre Stock about the new model – affectionately known as “le Chonk” – its specialized capabilities and extreme compute efficiency. Here’s what to know. Key capabilities of Mistral Large 4 Lample and Stock said that Mistral Large 4 was built to process both text and images, closing performance gaps from previous Mistral models. Specifically, Large 4 touts more advanced cybersecurity and agentic capabilities, as well as a better ability to handle manufacturing- and finance-related tasks. Cybersecurity Mistral Large 4 was designed to excel at defensive cybersecurity tasks, including log inspection, code auditing and checking for newly published exploits or vulnerabilities. Beyond defensive measures, Mistral says the new model offers offensive cybersecurity measures to identify and address weaknesses. Agentic AI “Really, the core competency we’re looking at for ML4 is agentic capabilities,” said Stock. Mistral Large 4 was built with a focus on agentic https://www.cnet.com/tech/services-and-software/what-is-agentic-ai-everything-to-know-about-artificial-intelligence-agents/ performance for planning, research or handling complex enterprise tasks that would take human experts hours or days to complete. The Mistral AI team said the model reaches “state-of-the-art” performance on critical agentic benchmarks like SWE-bench and Terminal Bench – evaluation frameworks for testing how well agents handle complex tasks – for tasks like coding, analyzing and pulling requests. Geospatial and grounding capabilities The Mistral team says the model outperforms competitors in aerial and satellite imagery analysis, with the capacity to identify visual patterns and objects on maps, such as forest fires https://www.cnet.com/science/climate/wildfires-are-raging-across-the-world-how-nasa-is-using-satellite-data-to-fight-back/ . “On grounding capabilities, it outperforms all existing models, including the closed ones,” the press release for the model says. “These are critical as cyberattacks get faster and cheaper, companies write more code than ever and the world needs more chips to keep up with AI demand.” Grounding is a technique used in AI development to try to eliminate hallucinations https://www.cnet.com/tech/services-and-software/what-are-ai-hallucinations-why-chatbots-make-things-up-and-what-you-need-to-know/ – when the LLM generates inaccurate information – by anchoring the AI’s outputs in external resources, like documents and databases, rather than just the materials it was trained on internally. Grounding also helps keep data and information fresh while providing context for the AI’s responses. Mistral says the new model was also trained on more than 160 languages, including every official language of the European Union, as well as Latin and non-Latin languages, to answer prompts natively. Mistral will roll out ML4 on Oct. 27. Until then, the model is available for a public preview period, allowing the Mistral team to further assess it with feedback from developers, cybersecurity experts and state officials. The team will also use this period for additional reinforcement learning and fine-tuning. You can try the preview API here https://docs.mistral.ai/models/mistral-large-4-0 .