AGC-Bench - sdat — leaderboard

Metric: Dataset z-score. Source: huggingface.co. 82 models tracked.

Top models

#ModelScore
1Claude Opus 4.60.68
2Llama 4 Scout0.67
3Gemma 3 4B (IT)0.67
4Claude Opus 4.50.62
5Llama 3.3 70B Instruct0.62
6Claude Sonnet 40.61
7Claude 3.7 Sonnet0.58
8Llama 3.1 70B Instruct0.57
9Claude Sonnet 4.50.56
10Llama 3.2 11B Instruct0.56
11Llama 3.2 3B Instruct0.54
12Gemma 2 27B (IT)0.53
13Claude Opus 4.70.51
14Llama 3 8B Instruct0.51
15Llama 3.1 8B Instruct0.51

Interactive version: theaggregate.ai/benchmark?slug=agc-bench-sdat · How the rankings work · Data refreshed daily, snapshot 2026-07-22.