French firm Mistral made its trillion-parameter AI model, Mistral Large 4, publicly available. The company said this tool helps businesses and governments build sovereign cyber defense. Pierre Stock said, "Businesses and governments can now access frontier intelligence to defend against cyberattacks." Users can now download the model on Hugging Face.

French artificial intelligence (AI) firm Mistral has made its trillion-parameter open-source AI model, Mistral Large 4 (ML4), publicly available. The move comes as a major step in the ongoing competition between open-weight AI developers like Mistral and closed-model providers such as OpenAI and Anthropic. The company believes that businesses and governments shouldn't have to depend on external AI companies for advanced capabilities, especially in cybersecurity where data control is paramount.

Mistral's CEO, Arthur Mensch, stressed that AI should be a tool for empowerment, not dominance. He believes that with ML4, Mistral can deliver state-of-the-art performance at a much lower cost than closed labs.

The release of the model's weights ensures other companies and government agencies can now use ML4 for their own purposes.

Mistral claims its model matches the performance of Chinese open-weight competitor GLM-5.3, especially in cyber work.

However, ML4 uses 47% fewer tokens per task, making it a more cost-effective solution than GLM-5.3.

The company also claims that the model outperforms all other open models for agentic coding tasks on Artificial Analysis's Coding Agent Index, an AI benchmarking site.

Mistral said that amid rising cyber threats, ML4 offers top-notch sovereign cyber defense capabilities to businesses and governments worldwide.

Enterprise and public sector institutions can now download, deploy, and own frontier AI to detect, analyze, and respond to threats across code and IT infrastructure, the company added.

The weights for ML4 are available on Hugging Face and through its API.

Pierre Stock, Mistral's Vice President of Science, told TechCrunch that the release of the model's weights gives organizations more control over their AI usage for self-defense.

He said businesses and governments can now access frontier intelligence to defend against cyberattacks.

Stock also emphasized that open-sourcing the model allows companies to deploy it on-premise in their own ecosystem.