List your company

Language Modeling Evaluation on General Language Evaluation Suite (test)

69.95AE ScoreROOT
57.688460.871764.05567.2383May 18, 2026

Evaluation results

MethodLinks
ROOTBackbone=Qwen3-1.7B2026.0569.9531.7487.725.257.6840.5325.871.7631.7754.9918.5916.2540.17
AMOBackbone=Qwen3-1.7B2026.0569.3632.688625.137.2841.0726.471.763456.221.131840.75
Muon-CANSBackbone=Qwen3-1.7B2026.0569.3232.0886.524.656.1841.042570.8433.7854.2217.2816.4639.78
Muon-PEBackbone=Qwen3-1.7B2026.0568.9831.6685.325.137.5440.9626.471.8232.4454.7820.8816.9340.24
AdamWBackbone=Qwen3-1.7B2026.0568.8130.3487.224.76.640.562571.7633.355.6418.6716.4639.92
Muon-KJBackbone=Qwen3-1.7B2026.0568.2231.7485.724.85.5440.7126.472.0933.785520.7216.4540.1
Muon-YouBackbone=Qwen3-1.7B2026.0567.5930.286.924.356.8740.4725.472.5234.0756.1219.2517.8940.14
AdaMuonBackbone=Qwen3-1.7B2026.0567.4231.4885.824.785.2439.6625.670.8933.6853.0420.3916.939.57
Muon-TurboBackbone=Qwen3-1.7B2026.0567.1731.7486.224.56.8139.9324.471.5533.0153.7520.816.5139.7
Muon-PEOptimizer=Muon-PE, Pre-training Budget=10x Chinchilla, Backbone=Qwen3-0.6B2026.0566.9832.088625.137.5437.832670.2332.4454.7821.291739.78
Muon-KJOptimizer=Muon-KJ, Pre-training Budget=10x Chinchilla, Backbone=Qwen3-0.6B2026.0566.2533.0386.225.356.8938.3325.469.732.7355.6418.8417.7339.67
Muon-PEOptimizer=Muon-PE, Pre-training Budget=5x Chinchilla, Backbone=Qwen3-0.6B2026.056627.385257.7737.5424.66932.3455.6720.6417.6539.04
AMOOptimizer=AMO, Pre-training Budget=10x Chinchilla, Backbone=Qwen3-0.6B2026.056633.7886.425.137.7238.7725.669.4833.4955.9320.6417.8240.06
Muon-KJOptimizer=Muon-KJ, Pre-training Budget=5x Chinchilla, Backbone=Qwen3-0.6B2026.0565.9128.7585.525.546.5837.2924.268.8832.7355.0119.6617.3238.95
AMOOptimizer=AMO, Pre-training Budget=5x Chinchilla, Backbone=Qwen3-0.6B2026.0565.7628.3385.725.027.7737.6125.669.2132.6755.2320.3717.9839.27
AMOBackbone=Qwen3-0.6B2026.0563.1727.1982.624.897.835.0421.668.6631.5353.0920.2317.3937.77
AMOOptimizer=AMO, Pre-training Budget=2x Chinchilla, Backbone=Qwen3-0.6B2026.0562.3727.2282.325.237.0235.742469.1531.5852.1720.3317.6537.9
Muon-CANSBackbone=Qwen3-0.6B2026.0562.3326.1980.324.097.9535.0422.266.9232.0652.7220.391737.27
Muon-PEOptimizer=Muon-PE, Pre-training Budget=2x Chinchilla, Backbone=Qwen3-0.6B2026.0562.2527.0382257.7735.0422.466.9232.0652.7219.7417.1937.51
Muon-KJOptimizer=Muon-KJ, Pre-training Budget=2x Chinchilla, Backbone=Qwen3-0.6B2026.0562.2127.0581.324.335.8435.8122.668.2831.252.9618.7617.5837.33
Muon-KJBackbone=Qwen3-0.6B2026.0562.1626.7981.923.87.334.7523.267.931.6751.931916.9237.28
Muon-PEBackbone=Qwen3-0.6B2026.0561.5326.7983.524.016.0934.7522.267.8531.9652.1720.4816.6437.33
Muon-TurboBackbone=Qwen3-0.6B2026.0561.4925.5182.723.456.435.0524.868.6630.7252.0919.8217.7137.37
Muon-YouBackbone=Qwen3-0.6B2026.0560.7727.2282.224.458.2234.622.867.0331.3953.041917.6537.36
ROOTBackbone=Qwen3-0.6B2026.0560.0125.68125.436.2633.9920.866.9729.7650.2821.717.6836.62
AdamWBackbone=Qwen3-0.6B2026.0558.9624.498223.516.7133.412166.2730.8150.1219.917.9936.26
AdaMuonBackbone=Qwen3-0.6B2026.0558.1624.9180.824.026.133.3322.266.9231.9652.4920.4817.3936.56