Timeline of Mistral Medium 3.1

Mistral Medium 3.1 scores 13.86% on Terminal-Bench

Mistral Medium 3.1 scores 0% on Terminal-Bench

Mistral Medium 3.1 scores 40.64% on τ²-bench

Mistral Medium 3.1 scores 54.16% on MMMU-Pro

Mistral Medium 3.1 scores 39.8% on IFBench

Mistral Medium 3.1 scores 32.41% on SciCode

Mistral Medium 3.1 scores 4.68% on Humanity's Last Exam

Mistral Medium 3.1 scores 58.79% on GPQA

Mistral Medium 3.1 scores 9.89 on Artificial Analysis Intelligence Index

Mistral Medium 3.1: context length changed from 131072 to 128000

Mistral AI La Plateforme lists Mistral Medium 3.1 at $0.2 in / $1 out per 1M tokens

Mistral AI La Plateforme lists Mistral Medium 3.1 at $0.4 in / $2 out per 1M tokens

New model: Mistral Medium 3.1 (Mistral AI)