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

Reasoning on GPQA Diamond (Score %)

85.5GPQA Diamond Reasoning Score

Qwen-3.5-27B

Updated 14d ago

Evaluation Results

Method	Links
Qwen-3.5-27B 2026.05		85.5
Gemma-4-31B 2026.05		84.3
Qwen-3.5-9B 2026.05		81.7
DeepSeek-R1 2026.05		71.5
Qwen-3-32B 2026.05		65.2
DeepSeek-V3 2026.05		59.1
Llama-3.3-70B 2026.05		50.5
Soft+ 2026.04		49.1
Qwen-2.5-72B 2026.05		49
Pair+ 2026.04		48.9
MPO+ 2026.04		48.5
DPO 2026.04		47.8
Llama-3.1-70B 2026.05		46.7
Mag+ 2026.04		46.6
Mistral-Small-3.2-24B 2026.05		46.1
GPT-4o-Mini 2026.05		40.2
MPO+ 2026.04		37.3
Mag+ 2026.04		36.9
Pair+ 2026.04		36.8
Soft+ 2026.04		35.4
Gemma-3-12B 2026.05		34.9
Base 2026.04		34.1
Qwen-2.5-7B 2026.05		33.8
DPO 2026.04		33.7
RFT 2026.04		33.2
Mixtral-8x22B 2026.05		33.2
Base 2026.04		32.1
Llama-3.1-8B 2026.05		30.4
RFT 2026.04		27.9
Soft+ 2026.04		26.7
DPO 2026.04		26.3
MPO+ 2026.04		25.5
Mag+ 2026.04		24.1
Pair+ 2026.04		23.6
Base 2026.04		22.3
RFT 2026.04		21.4