AidanBench: leaderboard
Open-ended idea generation: how many coherent, mutually novel answers a model can keep producing for the same question before repeating itself or losing coherence.
Metric: Novel Answers. Source: aidanbench.com. Saturation forecast: Estimated already saturated. 56 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 | 6054 | #183 |
| 2 | O3 Mini (High) | 4921 | #272 (O3 Mini) |
| 3 | O3 Mini (Medium) | 3368 | #272 (O3 Mini) |
| 4 | Gemini 2.5 Pro (Preview 03-25) | 3211 | #123 |
| 5 | GPT-4.5 (Preview) | 3003 | #244 |
| 6 | Claude 3.5 Sonnet | 2671 | #350 |
| 7 | O3 Mini (Low) | 2300 | #272 (O3 Mini) |
| 8 | Grok 3 Mini Beta | 2284 | #296 |
| 9 | Claude 3.7 Sonnet (Thinking) | 2170 | #249 (Claude 3.7 Sonnet) |
| 10 | Gemma 3 27B (IT) | 2028 | #523 |
| 11 | Claude 3.7 Sonnet | 1906 | #249 |
| 12 | O1 Preview | 1875 | #193 |
| 13 | Gemini 2.0 Flash (Preview) | 1872 | #370 |
| 14 | Grok 3 | 1795 | #304 |
| 15 | Gemma 3 12B (IT) | 1785 | #664 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aidanbench · How It Works · Data refreshed daily, snapshot 2026-10-20.