AGI-Eval Community Base - Learning (English): leaderboard

AGI-Eval Community private-set Base Language Models evaluation (Learning (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

#ModelScoreOverall rank
1DeepSeek V3 Base91.8
2DeepSeek-V3.1-Base91.69
3Kimi K2 Base91.54#939
4GLM-4.5-Base90.94
5DeepSeek-V3.2-Exp-Base89.7
6Llama 3.1 405B Base84.64
7Qwen 2.5 32B (Base)80.72
8Qwen3-4B-Base80.53#2439
9Qwen3-8B-Base80.5#1754
10Llama 4 Maverick Base79.65
11Qwen3-14B-Base79.14#1473
12ERNIE-4.5-300B-A47B-Base78.93
13Llama 3.1 70B Base76.88
14Qwen 2.5 7B (Base)76.08
15Qwen3-30B-A3B-Base76.07#1475

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