TipEurope’s Mistral releases its largest AI model for free: what it means for businesses that want AI and data in Europe
French company Mistral has made its largest model, Mistral Large 4, free to download. It was trained in Europe and can run on your own servers, so data does not have to leave the company. Through Mistral’s API it costs 0.70 dollars per million tokens in and 2.10 dollars out. Here is how to think about European AI.
Mistral is a French AI company and one of the few in Europe building large language models, the kind of AI behind bots like ChatGPT. On 29 October Mistral released the weights of Mistral Large 4, nicknamed Le Chonk. The weights are the model itself. Open weights mean anyone can download the model from Hugging Face and run it on their own computers.
The model has one trillion parameters and can read both text and images. Mistral says it was trained from scratch on European infrastructure and handles more than 160 languages. We have not seen an independent test of how good it is in the Nordic languages. According to Mistral it is the best open model for coding, with a special focus on IT security, finding and fixing security holes. These are Mistral’s own measurements.
Why does it matter? For public bodies and businesses that do not want to send their data to American or Chinese AI services, a strong European model they can run themselves is a new option. This becomes relevant as the public sector sets stricter limits on which AI services may be used. You do need powerful servers to run a model this large yourself. It is not something you run on an ordinary computer.
Price: through Mistral’s API, Mistral Large 4 costs 0.70 US dollars per million tokens in and 2.10 dollars per million tokens out, roughly 0.63 and 1.89 euros at the ECB rate of 29 October. A token is a piece of a word. The model is also available in Mistral Studio, on OpenRouter and in Microsoft Foundry. The licence lets you use, modify and sell services built on the model. The only exception is companies that sell the model itself as a service, which need a separate licence from Mistral.
Good to know: this is mainly news for businesses with their own IT or developers. Nothing changes for ordinary users today, and Mistral does not mention its chat app Le Chat in the launch. To find out whether your task needs AI that runs locally or whether a cloud service is enough, use the prompt Local AI or cloud AI in the Prompt Workshop.
How to get started
- First decide what the AI should do and how sensitive the information is. The prompt Local AI or cloud AI helps you.
- To test without your own servers: try the model in Mistral Studio or via OpenRouter with made-up or anonymised data.
- Compare the answers with the AI you use today, on real tasks and in your own language.
- If you are considering running it yourselves: ask your IT provider to cost servers, operations and licence before you decide.
More on TipAI
Sources
- Mistral: Mistral Large 4 (Le Chonk) is now open weight (29 okt 2026) (retrieved 2026-10-30)
- TechCrunch: Mistral releases weights for its trillion-parameter AI model, targeting cybersecurity (29 okt 2026) (retrieved 2026-10-30)
- ECB: referenskurser 29 okt 2026 (retrieved 2026-10-30)
Free cheat sheet: Get started with AI
A printable page with the basic formula, five things to try and how to use AI safely. Enter your email and we will send it straight away – free, unsubscribe any time.
Get the cheat sheet free →TipAI explains AI in plain language for beginners, families and small businesses. Every article has sources and a "How to get started" section.