Artificial Analysis launches AA-PersonalityBench: 30 personality traits and 21 communication styles across LLMs
ArtificialAnlys · x · 2026-10-20
Artificial Analysis released AA-PersonalityBench, benchmarking LLM personality via 30 traits (rolled into 8 indexes) and 21 programmatic communication characteristics, using 900 curated questions and LLM-judge Elo comparisons with no system prompt.
Top models per index: Expressiveness — Claude Fable 5.1, GLM-5.3; Organization — DeepSeek V4.1 Flash, Step 5 Preview; Informativeness — Muse Spark 1.3, Claude Fable 5.1, Claude Opus 5.5; Transparency — Claude Sonnet 5.5, GPT-6 Astra; Warmth — Muse Wonder; Conviction — Claude Fable 5.1; Imagination — Muse Spark 1.3; highest Slop — Llama 4 Maverick, GPT-4o, K2 Horizon.
Outliers in communication style: Qwen3.8 27B averages 1,030 words per response (4x GPT-6 Luna's 270); Nemotron 3 Ultra uses nearly 3 tables per response; GLM-5.3 averages 7 em dashes while Claude models almost never use one; K2 Horizon uses 2+ emojis per response.
Related event: Artificial Analysis Launches PersonalityBench to Rank Model Personalities(4 posts)→
More from Models
- DeepSeek V4.1 Flash ships a two-stage refusal guardrail, uncensoring effort finds — cephaloform · 2026-10-20
- Meta releases Muse Glimmer, a 30B open-weight agentic model that runs on a single consumer GPU — AIatMeta · 2026-10-20
- Maxwell debuts as 'world's first Code Execution Model', early access open — tri_dao · 2026-10-20
- d1-ultra tops Artificial Analysis Decision Index by a wide margin — JosephJacks_ · 2026-10-20
- Deutsche Bank: Meta's Muse still leads ChatGPT in downloads and DAUs 36 days post-launch — alexandr_wang · 2026-10-20
- Uncensored DeepSeek V4.1 Flash FP8 finetune trends on Hugging Face — dealignai · 2026-10-20