List your company

Mathematical Reasoning on GSM8K (PV, Best, Alpha Metrics)

97.7Ac (alpha=0.10)Oracle
56.72467.3627888.638May 13, 2026

Evaluation results

MethodLinks
OracleModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0597.7—97.710097.7100
OracleModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0595.6—95.610095.6100
R-Q8BModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0592.491.484.898.595.289.5
MVModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0591.490.1—97.295.279.2
R-L8BModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0591.490.783.898.795.285.2
J-Q30BModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0591.390.268.897.494.881.1
Ppl-StdModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.059188.87694.194.577.4
R-Q8BModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0590.988.388.395.495.181.8
J-G26BModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0590.988.767.595.294.977.4
PplModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0590.588.27894.794.580.5
R-L8BModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0590.487.384.294.49579.7
OracleModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0589.9—89.910089.9100
J-G26BModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.785.282.99294.375
MVModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.682.2—8494.264.7
PplModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.682.677.987.394.468.8
J-R1Q14BModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0589.684.764.188.89471.2
J-Q30BModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.483.870.388.294.468.7
Ppl-StdModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.481.572.584.293.965.9
J-R1Q14BModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0589.382.878.286.494.169.5
MVModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0588.873—71.392.961.7
R-Q8BModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0587.775.769.777.792.964.5
J-Q30BModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0587.773.659.173.592.462.5
R-L8BModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0587.27465.574.792.862.9
J-R1Q14BModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0587.273.761.274.892.362.2
J-G26BModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0586.972.359.974.192.562.1
PplModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0586.673.364.975.892.361.7
Ppl-StdModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0586.169.857.470.291.256.1
GreedyModel=Gemma-4-E4B, n_cal=200, m=16, splits=202026.0568——10068100
GreedyModel=R1-Qwen-1.5B, n_cal=200, m=16, splits=202026.0562.4——10062.4100
GreedyModel=Qwen3-1.7B, n_cal=200, m=16, splits=202026.0558.3——10058.3100