AGI-Eval Community Base - Coding Reasoning: leaderboard

AGI-Eval Community private-set Base Language Models evaluation (Coding Reasoning) 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
1Kimi K2 Base51.88#939
2Llama 4 Maverick Base51.44
3DeepSeek-V3.1-Base50.76
4Qwen3-30B-A3B-Base50.06#1475
5DeepSeek-V3.2-Exp-Base49.69
6DeepSeek V3 Base49.63
7Qwen3-14B-Base47.59#1473
8Qwen2.5-72B-Base46.36
9Qwen 2.5 32B (Base)44.97
10Qwen3-8B-Base44.67#1754
11Llama 3.1 405B Base42.37
12Qwen2.5-14B-Base42.08
13Qwen3-4B-Base40.75#2439
14Qwen 2.5 7B (Base)40.26
15ERNIE-4.5-300B-A47B-Base39.17

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