ModelsSelected

AA-PersonalityBench:Kimi K3.1 Pro 性格特征相比前代变化最大

Why we picked it

Moonshot 新的 Kimi K3.1 Pro 在 30 项性格测试里平均挪了 30 分,比 GPT-6 升级的 23 分还猛,讨论模型调性时可以看看这份数据。

Artificial Analysis 的 AA-PersonalityBench 用 30 项人格特质对模型做 1 到 100 的百分位排名。Kimi K3.1 Pro 相比 Kimi K3 在全部 30 项特质上平均移动了 30 个百分点,其中 18 项特质的变动幅度达到 25 个百分点以上。这是该基准记录到的模型换代间最大性格变化,此前最大的是 GPT-6 Sol 到 GPT-6.1 Sol 的 23 分。

Original · Artificial Analysis

In AA-PersonalityBench, Kimi K3.1 Pro shows the biggest personality change we’ve seen compared to a direct predecessor

Each model is percentile ranked from 1 to 100 on each of 30 personality traits. Across all 30 traits, Kimi K3.1 Pro's ranks moved by an average of 30 points from Kimi K3's (next largest is 23, for GPT-6 Sol to GPT-6.1 Sol), and 18 of the 30 traits moved by 25 points or more.