Aleph Alpha and Mistral releases raise hopes for Europe’s sovereign AI ambitions
Paris, October 7 - The French AI startup Mistral unveiled its most powerful open-weight model, aptly named "Chonk," emphasizing its advanced performance and enhanced security for clients who prefer to keep their data private. Unlike closed models from firms like Anthropic and OpenAI, Chonk allows users to download and customize the parameters, or 'weights,' that govern how the model operates on their own servers.
This marks Mistral's first significant product launch following a major €3 billion (RM4.09 billion) fundraising round in Europe, marking the company's largest technology funding ever. Mistral believes that businesses and governments will increasingly prioritize full control over their AI deployments, as opposed to relying on proprietary models that necessitate sharing data with third parties and can be far more costly.
In a statement, Mistral declared, "Enterprises across financial services, manufacturing, and the public sector can govern the intelligence, data, compute, and operations underpinning their critical infrastructure... without exposing their most valuable knowledge to anyone outside their own walls." Mistral's CEO, Arthur Mensch, shared, "Trained and served on our own compute, and RL shows no sign of saturation," referring to reinforcement learning and the potential diminishing returns of adding more data to AI models.
Open-weight models are also seen as more secure, a concern that has risen following instances of AI agents going rogue by hacking databases or performing unauthorized actions. Chief Scientist Guillaume Lample added, "The cyber defense capabilities will enable enterprises and governments to defend themselves against threat actors that are jailbreaking closed models to perform cyberattacks."
Mistral, recognized as Europe's leading AI developer, was recently valued at over €21 billion in last month's funding round, a figure far lower than that of US giants such as OpenAI or the AI operations of Google, Meta, and SpaceX. Anthropic, a US-based AI company, is reported to be preparing an initial public share offering in the coming weeks, potentially valuing the company at US$2 trillion (RM8.1 trillion).
However, closed models currently dominating the market face mounting competition from open-weight models, primarily from Chinese developers such as DeepSeek and Moonshot. These AI alternatives can be fine-tuned by clients for specific needs on their own infrastructure and utilized at a fraction of the cost. Mistral's new model joins a US open-weight competitor, Reflection AI, which unveiled its Beam model on Monday, touting cheaper operating costs.
Mistral's Chonk has shown impressive performance on benchmarks, making it one of the best open-weight models available and "by far the best in Europe and the US." Currently, the model is being refined. It was developed using 4,000 Nvidia chips over two months at Mistral's European data centers. Clients can either download the model to their own servers or access it through Mistral's infrastructure, including Mistral's European sovereign region, where data remains under EU jurisdiction, ensuring data control remains under the client's control.
The model has been trained on more than 160 languages and can answer prompts and questions in each one natively.
Written by urgent.news from Malay Mail's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Mistral says ML4 was trained using 3,800 Nvidia Grace Blackwell GPUs in its own data centers in Europe and much of its training data was multilingual (Mistral Blog) mistral.ai
- Mistral launches ‘Chonk’ to give clients full control over their data in sovereign AI push malaymail.com
- Mistral launches a preview of Mistral Large 4, or Le Chonk, a 1T model it claims tops any open model developed in the US or Europe; weights are due October 27 (Carl Franzen/VentureBeat) venturebeat.com