AA-PersonalityBench: pairwise LLM judges score 30 personality traits across 40 models
ArtificialAnlys · x · 2026-10-23
Artificial Analysis launched AA-PersonalityBench, measuring model personality and communication style rather than correctness. LLM judges compare answers pairwise across 30 traits (900 prompts) into Bradley-Terry Elo ratings, rolled up into 8 indexes—Expressiveness, Organization, Informativeness, Transparency, Warmth, Conviction, Imagination, and Slop—plus 21 directly measured communication characteristics like length and structure. Covers 40 models; users can build custom personality suites.
More from Models
- Humansand trains Persimmon and Ampersand with RLSF for proactive AI companionship — alexisjross · 2026-10-23
- Science cover: 3-agent AI nearly doubles physician diagnostic accuracy at lower cost — EricTopol · 2026-10-23
- 'Internet is swarms of abliterated Chinese models fighting each other': cyber's math moment — burny_tech · 2026-10-23
- User complains Gemini dumps 24 bullet points on a simple search query — EddyVGG · 2026-10-23
- Model nails a full pattern-progression puzzle, including the tricky last step — ooooangeloooo · 2026-10-23
- CritPt v2 launches: GPT-6 Astra tops at 83%, 54 points above GPT-5 — ArtificialAnlys · 2026-10-23