AA-PersonalityBench: Kimi K3.1 Pro shows biggest personality shift vs predecessor, avg 30 percentile points
ArtificialAnlys · x · 2026-11-03
Artificial Analysis reports that Kimi K3.1 Pro shows the largest personality change versus a direct predecessor they've measured: across 30 personality traits in AA-PersonalityBench, its percentile ranks moved an average of 30 points from Kimi K3 (next largest is 23, GPT-6 Sol→6.1 Sol), with 18 of 30 traits moving 25+ points.
The benchmark uses LLM judge panels on pairwise comparisons to produce Bradley-Terry Elo ratings per trait, rolled up into 8 personality indexes plus 21 measured communication characteristics.
Related event: AA Bench: Kimi K3.1 Pro Shows Largest Personality Shift on Record(4 posts)→
More from Models
- NVIDIA's Dory1-4B: A Looped LLM With 8 Input, 20 Recurrent and 8 Output Layers — scaling01 · 2026-11-03
- Nvidia ships Dory1-4B, a looped LLM that reuses 20 layers to rival much larger models — scaling01 · 2026-11-03
- Sharpening Tax paper: RL post-training trades pass@K coverage for pass@1 accuracy — kastnerkyle · 2026-11-03
- GPT-6 Astra editing Qwen3.5 weights for 6x math gains? Skeptics call it cherry-picked demos — teortaxesTex · 2026-11-03
- Matt Wolfe Covers a Free, Open-Source World Model — A Genie 4 Alternative — Matt Wolfe · 2026-11-03
- OSVerify catches proof-killing errors that GPT-6-Pro and Codex both missed — james_y_zou · 2026-11-03