Paris-based startup Mistral unveiled its latest artificial intelligence model, ML4, on Tuesday, emphasizing enhanced performance and security features designed to address growing concerns over data privacy. The new open-weight model, named Chonk, allows users to download and modify the underlying parameters, enabling tailored AI deployments without relying on proprietary, closed systems.

This release marks Mistral’s first major product launch since it secured a record-breaking fundraising round last month, raising three billion euros ($3.4 billion) — Europe’s largest technology investment to date. Following this round, Mistral’s valuation surpassed 21 billion euros, positioning it as a leading AI company on the continent, though still significantly lower than some U.S.-based competitors in the sector.

Mistral has positioned Chonk as a solution for enterprises and government entities that require direct control over AI operations to safeguard sensitive information. The company highlighted potential use cases across various sectors, including financial services, manufacturing, and public institutions, where governing data and computational processes internally is critical. The startup’s approach contrasts with the dominant models provided by firms like Anthropic and OpenAI, whose closed architectures typically demand sharing data with external servers and may carry higher costs.

The Chonk model is currently accessible as a public preview aimed at developers and governmental bodies for evaluation purposes. Full access to the model’s parameters, or "weights," is scheduled for release on October 27. This openness aligns with a broader trend of open-weight AI models gaining traction, particularly those developed by Chinese firms, which offer the advantage of customization and reduced operational expenses compared to closed alternatives.

In the wider market, companies such as Anthropic are preparing for major milestones, including an anticipated initial public offering. Reports indicate that Anthropic’s valuation could reach around $2 trillion, reflecting the significant investor interest in AI technologies. Nevertheless, the competition from open-weight models like Mistral’s signifies a shift in the industry landscape, potentially challenging the dominance of proprietary AI providers that have thus far shaped the market dynamics.