List your company

Medical Question Answering on Medical Benchmarks (MedQA, MedMCQA, BULLET) (test)

0.5533MedQA AccuracyC3oT
0.2690680.3428590.416650.490441Feb 2, 2026

Evaluation results

MethodLinks
C3oTBackbone=Qwen2.5-7B2026.020.55330.6150.38330.5172430.1297.32429.26385.561.42
CoT-ValveBackbone=Qwen2.5-7B2026.020.550.610.40330.5211564.33411.82584.32520.161.17
S3-CoTBackbone=Qwen2.5-7B2026.020.5450.60170.34830.4983302.19192.57309.47268.071.45
S3-CoTscBackbone=Qwen2.5-7B2026.020.52670.59330.33830.4861304.07195.01306.22268.431.3
Efficient_pBackbone=Qwen2.5-7B2026.020.49170.60330.31670.4706118.4362.23107.1195.921.5
TokenSKIPBackbone=Qwen2.5-7B2026.020.45670.51330.23170.4006461.66335.16464.99420.6-0.21
Standard_pBackbone=Qwen2.5-7B2026.020.40830.560.26830.4122498.3345.58514.86452.91—
S3-CoTscBackbone=DeepSeek-R1-7B2026.020.40.3850.28170.35561,647.53960.441,555.831,387.930.21
AutoTHINKBackbone=DeepSeek-R1-7B2026.020.3950.38330.31330.36391,742.881,117.91,692.671,517.820.26
S3-CoTBackbone=DeepSeek-R1-7B2026.020.3950.36670.30170.35451,648.161,024.411,608.561,427.040.16
Standard_pBackbone=DeepSeek-R1-7B2026.020.38670.37330.30830.35611,865.661,339.952,169.361,791.66—
LASER_DEBackbone=DeepSeek-R1-7B2026.020.38330.37670.29330.35111,354.27868.281,286.061,163.540.21
ShorterBetterBackbone=DeepSeek-R1-7B2026.020.37170.360.30470.3454692.33357.91716.785890.37
EffReaBackbone=DeepSeek-R1-7B2026.020.370.36330.31830.35061,736.26948.041,651.761,445.350.04
Efficient_pBackbone=DeepSeek-R1-7B2026.020.36830.34330.2750.32891,362.7746.791,297.021,135.5-0.4
C3oTBackbone=DeepSeek-R1-7B2026.020.3650.36830.32670.35331,832.411,265.821,802.021,633.420.01
LC-R1Backbone=DeepSeek-R1-7B2026.020.35830.3750.29170.34171,310.58672.81,248.251,077.21-0.01
CoT-ValveBackbone=DeepSeek-R1-7B2026.020.280.30830.24670.27831,153.821,676.99918.821,249.88-1.88