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

Graduate-level Science Reasoning on GPQA (Score)

59.1Score

Qwen3-8B

Updated 22d ago

Evaluation Results

Method	Links
Qwen3-8B 2025.10		59.1
Qwen3-4B 2025.10		54.5
Ouro-2.6B-Thinking-R4 2025.10		52.7
Deepseek-Distill-Qwen-7B 2025.10		51
Ouro-1.4B-Thinking-R4 2025.10		45.5
Qwen3-1.7B 2025.10		34
Deepseek-Distill-Qwen-1.5B 2025.10		33.2
Nemotron-CC_ASI+ 2026.03		27.1
DCLM 2026.03		25.67
Fineweb-Edu 2026.03		24.51
Nemotron-CC 2026.03		24.37
Nemotron-CC_ASI 2026.03		23.93
Ultra-Fineweb 2026.03		23.04
GPT-o1 2025.11		0.78
DeepSeek–R1 2025.11		0.72
GPT-4o 2025.11		0.54
GPT-4o-Mini 2025.11		0.4
GPT-3.5-Turbo 2025.11		0.31
Qwen2.5-3B-Instruct 2025.11		0.3
Falcon-7B-Instruct 2025.11		0.25
LLaMA2-13B-Chat 2025.11		0.23