Compared with Kimi K3, K3.1 Pro improves on the Artificial Analysis Cyber Index by 23 points, from 41% to 64%. This is mainly driven by a 37-point gain on CyberGym-E2E-AA, from 58% to 95%, the highest score of any model we have tested.
Compared with Kimi K3, K3.1 Pro improves on the Artificial Analysis Cyber Index by 23 points, from 41% to 64%. This is mainly driven by a…
Compared with Kimi K3, K3.1 Pro improves on the Artificial Analysis Cyber Index by 23 points, from 41% to 64%. This is mainly driven by a 37-point gain on CyberGym-E2E-AA, from 58% to 95%, the highest score of any model we have tested.
Artificial Analysis · X
Content notes
Original publication and rights belong to the source.