AIIQ Composite IQ: leaderboard
AIIQ converts scores from public leaderboards into IQ style points, interpolating along a 70 to 160 ladder set per benchmark from its own estimates of human performance. The composite is the mean of six equally weighted dimensions spanning reasoning, computer use and reliability, with missing ones imputed. This board reads it via benchmarklist.com, and higher is better. Source: aiiq.org/methodology/.
Metric: Composite IQ (imported). Source: www.aiiq.org. Source: benchmarklist.com. Saturation forecast: Around September 2027. 177 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 141 | #3 |
| 2 | GPT-6.1 Sol | 141 | #7 |
| 3 | GPT-6 Astra | 140 | #6 |
| 4 | Gemini 4 Argon | 140 | #8 |
| 5 | GPT-6 Sol | 137 | #14 |
| 6 | Claude Fable 5.1 | 137 | #9 |
| 7 | Claude Sonnet 5.5 | 137 | #11 |
| 8 | Claude Opus 5 | 136 | #12 |
| 9 | GPT-5.6 Sol | 134 | #19 |
| 10 | Claude Fable 5 | 133 | #16 |
| 11 | Claude Opus 4.7 | 132 | #49 |
| 12 | Claude Opus 4.6 | 131 | #63 |
| 13 | Muse Wonder Full | 131 | #10 |
| 14 | GPT-5.6 Terra | 130 | #41 |
| 15 | Gemini 3.8 Flash | 130 | #21 |
Interactive version: theaggregate.ai/benchmark?slug=aiiq-composite-iq · How It Works · Data refreshed daily, snapshot 2026-10-20.