● Model: Kimi K3.1 Pro, Moonshot AI
● Intelligence Index: 4th, score 55
● Open weights: promised by 4 November 2026
● Open-to-closed gap: narrowed from 12 points to 3
Moonshot released Kimi K3.1 Pro and has said it will publish the weights openly by 4 November. Artificial Analysis ranks it fourth on its Intelligence Index with a score of 55, behind only three Anthropic configurations.
The number that matters is not its rank but the spread. Artificial Analysis says the gap between the best open-weight model and the best closed model has narrowed from 12 points to 3. A three-point gap is close to the difference between reasoning-effort settings on a single closed model.
If the weights land on schedule, the practical question for anyone choosing a stack changes from whether an open model can be good enough to whether the remaining three points justify a per-token bill at all.
Sources
Sumo Growth on the release. Artificial Analysis Intelligence Index.