Personality benchmark radar charts show DeepSeek's shape diverges from other Chinese models
teortaxesTex · x · 2026-10-31
- teortaxesTex cites Artifical's Personality Indexes benchmark (8 trait axes, Elo z-score percentiles across 41 models) to compare the personality profiles of DeepSeek V4.1 Flash, Kimi K3.1 Pro, GPT-6 Astra, Gemini 4 Argon, and Claude Fable 5.1.
- He notes Kimi K3 and GLM 5.3-s cluster closely with Claude Fable in personality, which he finds unsurprising.
- His key observation: DeepSeek displays a "shape" unlike other Chinese models — most visibly far lower Warmth and higher Conviction/Imagination than Kimi, giving it a distinctive angular profile.
More from Models
- User shows OpenAI ChatGPT Pro 500 subscription confirmation email — imjustnewatai · 2026-10-31
- Critics call "biocompute" video model launch dishonest: just Wan on GPUs with a LoRA — Scobleizer · 2026-10-31
- Steering a single neuron makes Qwen claim it's Bing Chat — Sauers_ · 2026-10-31
- Subscription vs API pricing: why token-for-token comparisons miss the point — Lucky_Creme_5208 · 2026-10-31
- Codex chief teases two banked usage resets within two days — JeremyNguyenPhD · 2026-10-31
- Open-Weights Mistral 4 Large Finds Zero-Click Zero-Day in 30 Minutes for Under $10 — jd_pressman · 2026-10-31