List your company

Science Question Answering on GPQA (Acc, Peak, Dep)

38.22AccuracyTokenSkip
19.479224.344629.2134.0754Apr 4, 2026

Evaluation results

MethodLinks
TokenSkipModel Series=Qwen2.5-7B, Budget setting=true2026.0438.225,61121.2
LThinker++Model Series=Qwen2.5-7B, Budget setting=true2026.0438.223,26516.1
LThinker++Model Series=Llama3.1-8B, Budget setting=true2026.0433.674,10122.9
VanillaModel Series=Qwen2.5-7B, Budget setting=true2026.0432.495,49320
VanillaModel Series=Llama3.1-8B, Budget setting=true2026.0431.825,77321.7
LThinker*Model Series=Qwen2.5-7B, Budget setting=true2026.0428.963,49623.2
LThinker*Model Series=Llama3.1-8B, Budget setting=true2026.0428.284,27134.5
CoTModel Series=Qwen2.5-7B, Budget setting=true2026.0426.769680.5
TokenSkipModel Series=Llama3.1-8B, Budget setting=true2026.0426.66,01623.8
Distill-R1Model Series=Qwen2.5-7B, Budget setting=true2026.0424.756,71831
CoTModel Series=Llama3.1-8B, Budget setting=true2026.0424.751,2310.9
Distill-R1Model Series=Llama3.1-8B, Budget setting=true2026.0420.23,97216