Share your thoughts, 1 month free Claude Pro on usSee more

Language Understanding on MMLU-Pro (Score)

81.4MMLU-Pro Score

GLM-4.5-Air

Updated 18d ago

Evaluation Results

Method	Links
GLM-4.5-Air 2026.01		81.4
Solar Open 2026.01		80.4
gpt-oss-120b 2026.01		80.4
gpt-oss-120b 2026.01		78.6
Baseline 2026.07		69
Trinity Large Base 2026.02		66.02
AFM-k5984d3s (Ours) 2026.07		65.9
Uniform schedule 2026.06		54.21
Linear schedule 2026.06		49.05
Cosine schedule 2026.06		46.39
Qwen2-7B 2024.07		40
Llama-3-8B 2024.07		35.4
Uniform schedule 2026.06		33.94
Gemma-7B 2024.07		33.7
Polynomial schedule 2026.06		33.16
Mistral-7B 2024.07		30.9
Qwen1.5-7B 2024.07		29.9
Linear schedule 2026.06		24.97
AR 2026.04		24.7
Exponential schedule 2026.06		21.41
Cosine schedule 2026.06		21.26
TIDE-Cross 2026.04		14.52
TIDE-Shared 2026.04		14.48
TIDE-Shared 2026.04		14.02
Uniform schedule 2026.06		13.95
BD3LM 2026.04		13.8
Linear schedule 2026.06		13.58
CALM 2026.04		13.47
Polynomial schedule 2026.06		13.38
KL 2026.04		13.19
TIDE-Cross 2026.04		12.88
Cosine schedule 2026.06		12.57
Exponential schedule 2026.06		11.83
Polynomial schedule 2026.06		11.04
Exponential schedule 2026.06		10.52
Cosine schedule 2026.06		10.06
Polynomial schedule 2026.06		9.61
Polynomial schedule 2026.06		9.23
Linear schedule 2026.06		9.16
Exponential schedule 2026.06		8.99
Exponential schedule 2026.06		8.63
Uniform schedule 2026.06		8.44
Linear schedule 2026.06		5.92
Cosine schedule 2026.06		2.27