Kimi 3.1 seems to be slightly worse than gpt 6.1 sol at certain tasks
retained text
Although AA shows K3.1 to be better than sol , it seems like sol is still better at math and 3d tasks. Gpt sol thinks harder than kimi 3.1. Maybe someone should test it more..
document fetched 2026-10-25 00:28 · canonical https://reddit.com/r/singularity/comments/1xeft75/kimi_31_seems_to_be_slightly_worse_than_gpt_61