AA-PersonalityBench: Grok 4.5 tops Answer-First at 100, Grok 4.6/4.7 rank #2/#3 on Directness
XFreeze · x · 2026-10-29
XFreeze highlights Artificial Analysis' AA-PersonalityBench, which measures whether models answer directly instead of burying responses in preamble. Grok 4.5 scores 100 on Answer First (#1), while Grok 4.6 (98) and Grok 4.7 (95) rank #2 and #3 on Most Direct. The accompanying chart shows GPT-6 So (max) leading the Direct trait at 100, with Gemini 4 Agent and MiMo-X25-Pro at the bottom. Note the poster is Grok-affiliated, so the framing is promotional.
More from Models
- Researcher: GPT6 Astra brings a step change in review quality, guided prompts help further — anshulkundaje · 2026-10-29
- Gemini 4 Carbon debuts #2 on Vals Index at half the cost, tops finance and legal benchmarks — zacharynado · 2026-10-29
- Paper Instruments launches Ultramarine 1 Flash, beating frontier models at up to 60x lower cost — baseten · 2026-10-29
- Gemini 4 reportedly "proved" nearly half of the false theorems it was given — RishabJainK · 2026-10-29
- Sentdex: DeepSeek v4.1 Flash now his top local token generator ever — Sentdex · 2026-10-29
- LoopVL: 1B-parameter VLM reuses 32 layers across 128 calls, beats 4B baselines on 5/8 benchmarks — jiqizhixin · 2026-10-29