AGI-Eval Community Base - Interaction: leaderboard

AGI-Eval Community private-set Base Language Models evaluation (Interaction) 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

#ModelScoreOverall rank
1Qwen2.5-72B-Base81.21
2Qwen 2.5 32B (Base)80.78
3DeepSeek-V3.1-Base79.87
4Kimi K2 Base79.83#939
5DeepSeek V3 Base79.45
6DeepSeek-V3.2-Exp-Base78.38
7ERNIE-4.5-300B-A47B-Base78.3
8Qwen2.5-14B-Base78.16
9Llama 4 Maverick Base77.13
10Llama 3.1 405B Base77.12
11Qwen3-14B-Base76.91#1473
12Qwen3-8B-Base76.86#1754
13GLM4-9B-Base76.75
14GLM-4.5-Base75.96
15Llama 3.1 70B Base75.04

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-interaction · How It Works · Data refreshed daily, snapshot 2026-11-22.