MiMo-V2.6-Pro ends 34% of responses with a question, tops Artificial Analysis personality bench
ArtificialAnlys · x · 2026-11-06
Artificial Analysis's AA-PersonalityBench measures the share of model responses that end with a question, ranging from 34% down to 0%.
Most inquisitive: MiMo-V2.6-Pro (34%), GPT-6 Astra (32%), GPT-6.1 Sol (29%), Muse Glimmer (28%), Muse Spark Open (26%).
Least inquisitive (six models under 0.5%): GPT-5.4 mini and Nemotron 3 Ultra (0.0%), DeepSeek V4.1 Flash and its 1017 build (0.1%), GPT-3.5 Turbo (0.2%), Grok 4.7 (0.4%).
The data highlights how strongly models diverge in conversational personality — reflexive follow-up questions have become a signature habit of several newer models.
More from Models
- Intelligence now 'too cheap to meter': investor drops Codex for all-you-can-eat Claude at $100 — StewartalsopIII · 2026-11-06
- Menlo report: Enterprise AI spend soars to $177B, up 5.4x in a year as agents reshape buying — mmurph · 2026-11-06
- vLLM releases Kimi K3.1 Pro pruned 75%: 896→224 experts per MoE layer, ~1.5TB down to ~470GB — victormustar · 2026-11-06
- Tom Goldstein teases more sphere models coming very soon, with Meta AI collaborators — tomgoldsteincs · 2026-11-06
- GPT-6-Sol's unscored RCE chain: fixing a scorer bug ties it with GLM-5.3-Flash — teortaxesTex · 2026-11-06
- Same 10 prompts, one year apart: GPT-6 output dwarfs the GPT-5 3D editor demo — Just_Lingonberry_352 · 2026-11-06