IKP - T3: leaderboard
IKP's third obscurity tier tests less common factual knowledge, including researcher recognition. Accuracy on all 183 cleaned T3 questions, answered without retrieval or supplied evidence.
Metric: Accuracy (%). Source: 01.me. Saturation forecast: Estimated already saturated. 197 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 100 | #64 |
| 2 | GLM-5.2 | 100 | #80 |
| 3 | GPT-5 | 99.45 | #105 |
| 4 | GPT-4.1 | 99.45 | #258 |
| 5 | Claude Opus 4.6 | 99.45 | #67 |
| 6 | Claude Fable 5 | 99.45 | #20 |
| 7 | MiMo-V2.5-Pro | 99.45 | #167 |
| 8 | GPT-5.3 | 99.45 | #104 |
| 9 | GPT-5 Pro | 99.45 | #89 |
| 10 | Claude Opus 4.5 (High) | 99.45 | #87 (Claude Opus 4.5) |
| 11 | Gemini 3 Flash (Preview) (Medium) | 99.45 | #90 (Gemini 3 Flash (Preview)) |
| 12 | Gemini 2.5 Pro (Medium) | 99.45 | #159 (Gemini 2.5 Pro) |
| 13 | Gemini 2.5 Pro (Preview) | 99.45 | #136 |
| 14 | GPT-5.5 | 98.91 | #33 |
| 15 | DeepSeek V4 Flash | 98.91 | #155 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ikp-t3 · How It Works · Data refreshed daily, snapshot 2026-11-05.