AGI-Eval Community Base - Cognition (English): leaderboard
AGI-Eval Community private-set Base Language Models evaluation (Cognition (English)) measuring accuracy (%) across bilingual capability dimensions.
Metric: Accuracy (%). Source: agi-eval.org. Saturation forecast: Estimated already saturated. 25 models tracked. Updated 22 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek-V3.1-Base | 85.69 | |
| 2 | DeepSeek V3 Base | 84.83 | |
| 3 | GLM-4.5-Base | 83.18 | |
| 4 | DeepSeek-V3.2-Exp-Base | 83 | |
| 5 | Kimi K2 Base | 82.75 | #939 |
| 6 | Qwen 2.5 32B (Base) | 81.85 | |
| 7 | Qwen2.5-72B-Base | 79.46 | |
| 8 | Qwen3-30B-A3B-Base | 78.89 | #1475 |
| 9 | Llama 4 Maverick Base | 77.82 | |
| 10 | Llama 3.1 405B Base | 77.33 | |
| 11 | Qwen3-14B-Base | 77.29 | #1473 |
| 12 | Qwen2.5-14B-Base | 77.27 | |
| 13 | GLM4-9B-Base | 75.8 | |
| 14 | Llama 3.1 70B Base | 75.69 | |
| 15 | ERNIE-4.5-300B-A47B-Base | 75.27 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=agi-eval-community-base-cognition-english · How It Works · Data refreshed daily, snapshot 2026-11-22.