Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Model Performance Prediction on DeepSeek Model Families (Hold-out)

0.02MAE

TailoredBench

-0.49382.974356.44259.91065Feb 8, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.02
0.02
2026.02
0.156
2026.02
0.264
2026.02
0.73
2026.02
0.738
2026.02
0.926
2026.02
1.096
2026.02
1.392
2026.02
1.448
2026.02
1.535
2026.02
1.595
2026.02
1.681
2026.02
1.777
2026.02
1.873
2026.02
1.882
2026.02
1.95
2026.02
2.019
2026.02
2.083
2026.02
2.09
2026.02
2.132
2026.02
2.244
2026.02
2.304
2026.02
2.374
2026.02
2.459
2026.02
2.542
2026.02
2.624
2026.02
2.657
2026.02
2.816
2026.02
3.407
2026.02
3.7
2026.02
3.712
2026.02
3.74
2026.02
3.751
2026.02
3.81
2026.02
4.182
2026.02
4.216
2026.02
4.616
2026.02
4.616
2026.02
5.71
2026.02
6.195
2026.02
6.195
2026.02
6.464
2026.02
11.662
2026.02
12.865
2026.02
12.865