Mistral Large 4 scores 38 on the Intelligence Index, putting Europe back on the frontier map

As seen on the 24/7 Wall St. homepage on October 6, 2026.

Mistral plans to release the weights, which puts a frontier-adjacent European model into open hands and pressures the pricing power of closed US labs.

Mistral has released Mistral Large 4, scoring 38 on the Artificial Analysis Intelligence Index; France is back to having the most intelligent model from outside the US and China @MistralAI has released Mistral Large 4 in Research Public Preview, with plans to release the weights https://t.co/pzGQujfDlH
  • Replies5
  • Reposts13
  • Likes88
Continue ReadingShow less

Mistral Large 4 launched on October 6, 2026, in Research Public Preview, and its score of 38 on the Artificial Analysis Intelligence Index is enough to make it the most capable model built outside the United States and China. That is a meaningful threshold, because it means a European lab has clawed back a position at the top of the non-US, non-China tier after losing that ground to competitors.

The release arrives initially as a Research Public Preview, but Mistral has signaled plans to release the model weights publicly. Open weights change the competitive dynamics considerably: developers and enterprises can run the model themselves rather than paying per token, which directly pressures the pricing power of closed-source US labs that rely on API access fees as their primary revenue model.

France in particular has treated its AI sector as a matter of national industrial policy, and Mistral has been the flagship of that effort. Getting back to the top of the non-US, non-China rankings reinforces the case that European sovereign AI investment can produce frontier-adjacent results, a point that will matter to policymakers and enterprise buyers looking to diversify away from a handful of American providers.

The post came from Artificial Analysis, a benchmarking-focused account whose audience is technically specialized. Investors and developers tracking the open-weights AI space will want to follow how the full weight release is timed and what licensing terms accompany it, since those details will determine how broadly the model gets adopted.