TrackingAI IQ Test (Mensa Norway): leaderboard
Mensa Norway's public 35-question timed visual pattern puzzle test, verbalized for text-only models, measuring abstract reasoning on a standard IQ scale.
Metric: Mensa Norway Score (%). Source: trackingai.org. Saturation forecast: Around January 2027. 39 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 100 | #16 |
| 2 | Claude Opus 5 (Max) | 97.14 | #12 (Claude Opus 5) |
| 3 | GPT-5 Pro | 97.14 | #81 |
| 4 | Grok 4 | 94.29 | #181 |
| 5 | Gemini 3 Pro | 94.29 | #80 |
| 6 | O3 Pro | 94.29 | #116 |
| 7 | Claude Opus 5 | 91.43 | #12 |
| 8 | GPT-5 (Thinking) | 91.43 | #98 (GPT-5) |
| 9 | O3 | 88.57 | #124 |
| 10 | Claude Opus 4 (Thinking) | 88.57 | #163 (Claude Opus 4) |
| 11 | Grok 4.5 | 85.71 | #48 |
| 12 | O1 | 82.86 | #183 |
| 13 | Claude Sonnet 5 | 80 | #57 |
| 14 | Claude Opus 4 | 80 | #163 |
| 15 | GPT-5 | 77.14 | #98 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=trackingai-iq-test-mensa-norway · How It Works · Data refreshed daily, snapshot 2026-10-20.