Benchmarking Large Language Models for Graphem Leaderboard — The Aggregate

Benchmarking Large Language Models for Graphem — leaderboard

Metric: Direct (self-reported). Source: benchmarklist.com. 37 models tracked.

Top models

#	Model	Score
1	Qwen 3 8B	100.15
2	Qwen 3.5 2B	87.3
3	Qwen 2.5 7B	79.39
4	glm-4-9B	57.56
5	Gemma 3 4B	56.69
6	Qwen 3.5 4B	43.98
7	GLM-4 32B	35.62
8	Qwen 2.5 32B	35.48
9	Qwen 3 32B	32.46
10	Qwen 3 14B	32.11
11	Gemma 3 12B	28.77
12	Qwen 3.5 9B	27.47
13	Qwen 3.5 35B A3B	26.93
14	Gemma 2 27B	24.56
15	Llama 3.3 70B Instruct	23.48

Interactive version: theaggregate.ai/benchmark?slug=benchmarking-large-language-models-for-graphem · How the rankings work · Data refreshed daily, snapshot 2026-07-22.