Claude and GPT Hit Mensa Test Ceiling: Perfect Scores for 151 IQ, Beyond Human Scale
新智元 · wechat · 2026-09-26
Claude Fable 5.1 and GPT-6 Astra answered all 35 questions on the Mensa Norway IQ test perfectly, scoring 151 — above the 145 cap for humans — with perfect scores across their last 7 attempts.
Key facts:
- Maxim Lott's TrackingAI tests 30+ models weekly; GPT-5.6 Sol and Gemini 3.1 Pro hit 145, with Chinese models also in the top tier
- Only 2 AIs solved question 33; head models went from 64 (blind guessing) to 151 in 2.5 years, gaining 2.5 IQ points/month vs the Flynn effect's 3 points/decade
- To rule out training-data contamination, Lott created a never-published test; top models still neared its ceiling (GPT-5.6 Terra visual: 135/136), a year ahead of his prediction
- François Chollet remains skeptical via ARC benchmarks, estimating AI still 6 orders of magnitude behind humans in learning efficiency, though he now expects AGI before 2030
Related event: AI Maxes Out Mensa Norway IQ Test at 151, Beyond Human Ceiling(5 posts)→
More from Models
- DeepMind says Gemini agents mathematically verified Signal's crypto protocol in 4 weeks — Polymarket · 2026-11-11
- Leak: Google's upcoming model lineup includes Hydrogen, Krypton, Helix and a video model — koltregaskes · 2026-11-11
- Rumor: OpenAI's Astra 6.2 near, spotted in Codex logs and Google Cloud quotas — haider1 · 2026-11-11
- Same prompt head-to-head: Fable 5.2 spends 3x time, 2x cost for auteur flair — iamfakhrealam · 2026-11-11
- Krafton open-sources Raon-1 nano, a 1.5B on-device LLM distilled from MoE — DimitrisPapail · 2026-11-11
- Do typos discomfort LLMs? Tokenization quirk sparks debate — menhguin · 2026-11-11