KINA: leaderboard
KINA (Knowledge Index of Noah's Ark) — 899-item knowledge benchmark across 261 fine-grained disciplines with representativeness-aware sampling and tournament-audited annotation; overall accuracy.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2029. 42 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 53.17 | #58 |
| 2 | Claude Opus 4.6 | 49.92 | #63 |
| 3 | GPT-5.4 | 48.55 | #79 |
| 4 | Doubao-Seed-2.0-Pro-260215 | 44.99 | #77 |
| 5 | Gemini 3 Flash (Preview) | 43.91 | #82 |
| 6 | Qwen 3.5 397B A17B | 42.99 | #146 |
| 7 | Kimi K2.5 | 40.24 | #144 |
| 8 | GPT-5.2 | 39.52 | #110 |
| 9 | Qwen 3.5 27B | 39.35 | #187 |
| 10 | Qwen 3.5 122B A10B | 38.88 | #173 |
| 11 | DeepSeek V3.2 (Thinking) | 38.01 | #205 (DeepSeek V3.2) |
| 12 | Qwen 3 Max (2025-09-23) | 35.9 | #255 |
| 13 | GLM-5 | 35.85 | #141 |
| 14 | Qwen 3.5 35B A3B | 35.43 | #256 |
| 15 | Grok 4.1 Fast (Reasoning) | 33.73 | #217 (Grok 4.1 Fast) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=kina · How It Works · Data refreshed daily, snapshot 2026-10-20.