benchmark,split,n,accuracy,mean_tokens,truncated,hidden_reasoning humaneval,instant_r1,100,0.74,559.5,0.01,0.08 humaneval,thinking_off,100,0.83,1877.53,0.12,0.7 humaneval,thinking_on,100,0.82,2015.12,0.11,0.0 mbpp,instant_r1,100,0.65,251.83,0.01,0.04 mbpp,thinking_off,100,0.69,676.89,0.06,0.2 mbpp,thinking_on,100,0.8,1296.03,0.08,0.0 livecodebench,instant_r1,30,0.6,2385.2667,0.1,0.3 livecodebench,thinking_off,30,0.6667,3604.2333,0.1667,0.6 livecodebench,thinking_on,30,0.6,5382.7667,0.3667,0.0 gsm8k,instant_r1,100,0.88,439.13,0.02,0.1 gsm8k,thinking_off,100,0.88,563.4,0.02,0.19 gsm8k,thinking_on,100,0.85,789.05,0.03,0.0 ifeval,instant_r1,100,0.72,319.35,0.0,0.0 ifeval,thinking_off,100,0.75,437.04,0.02,0.05 ifeval,thinking_on,100,0.77,2327.91,0.19,0.0 ifbench,instant_r1,100,0.23,696.04,0.03,0.06 ifbench,thinking_off,100,0.27,1196.32,0.12,0.11 ifbench,thinking_on,100,0.34,3412.11,0.58,0.0 bfcl,instant_r1,70,0.7714,69.0,0.0,0.0 bfcl,thinking_off,70,0.7714,83.3857,0.0,0.0429 bfcl,thinking_on,70,0.9,357.0286,0.0,0.0