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

General Model Evaluation on Comprehensive Reasoning Suite (MMLU, HumanEval, GSM8K, GPQA, etc.)

69.3Average Score

CHORD

Updated 1mo ago

Evaluation Results

Method	Links
CHORD 2025.09		69.3
A-CHORD 2025.09		69.27
A-GRPO 2025.09		68.39
GRPO 2025.09		68.11
SFT → A-GRPO 2025.09		68.11
A-SFT → A-GRPO 2025.09		67.93
SFT → GRPO 2025.09		65.75
Qwen3-4B 2025.09		62.52
SFT 2025.09		60.8
A-SFT 2025.09		60.67
SFT → A-GRPO 2025.09		43.61
A-SFT → A-GRPO 2025.09		43.31
SFT → GRPO 2025.09		42.56
A-CHORD 2025.09		42.09
A-GRPO 2025.09		41.99
CHORD 2025.09		41.19
GRPO 2025.09		40.93
Qwen3-0.6B 2025.09		40.81
A-SFT 2025.09		39.83
SFT 2025.09		38.48