French firm Mistral has launched Le Chonk, a new 1 trillion-parameter AI model, claiming it can match the best from the US and China. Optimized for coding and cyberdefense, it’s designed to fill niches in manufacturing, finance and engineering. While US labs charge for their proprietary models, Mistral makes its open-source offering free. The company hopes this will eliminate hesitations for businesses to opt for open-source solutions.

The launch comes amid growing tensions over AI access, with the US government imposing restrictions on certain models. Mistral, with its recent $3.3 billion funding round, is positioning itself as a beacon of technological sovereignty in Europe. Its ability to provide a competitive, open-source alternative could change the dynamics of the AI race.

The stakes are high. As Andrea Renda, director of research at the Centre for European Policy Studies, points out, the EU’s push for technological independence, coupled with US hostility, has created an ideal scenario for Mistral. But the real question is whether open-source models can truly challenge the dominance of closed models, especially when it comes to critical security and defense tasks.

While Mistral is eager to show that businesses can benefit from owning their models, regardless of location, the success of Le Chonk will depend on how well it can compete in areas where proprietary models have traditionally excelled.