Researcher Warns Kimi K3 Max Fails Complex Statistical Auditing Tasks
Source Posts on X
A note of caution: I will say that when doing some complex statistical auditing of some of my prior academic work, Kimi K3 Max messed up in a bunch of ways, including misapplying statistics and applying some stuff badly. A bit from GPT-5.6 Pro critiquing K3 (which I agree with):
I think you really need to benchmark this in your area of interest, open weights models often have much more jagged frontiers in ways that are hard to spot in advance.
Posts from X
2 postsI think you really need to benchmark this in your area of interest, open weights models often have much more jagged frontiers in ways that are hard to spot in advance.
A note of caution: I will say that when doing some complex statistical auditing of some of my prior academic work, Kimi K3 Max messed up in a bunch of ways, including misapplying statistics and applying some stuff badly. A bit from GPT-5.6 Pro critiquing K3 (which I agree with):