List your company
ResearchFormerly WizWand

Test-Time Scaling Selection on RouterBench Quality-Priority

0.7924RewardOracle
0.5034880.5784940.65350.728506May 29, 2026
Updated 5mo ago

Evaluation results

MethodLinks
OracleNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.792468.12115.7
OracleBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.742659.1220.3
UNISCALEBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.64548.1226.6
UNISCALENumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.630657.371,374.7
TSNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.6243541,875.7
GreedyBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.618446.567.6
RandomNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.617552.881,359.8
MLPNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.605547.5353.4
NeuralUCBNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.592950.251,819.7
TSBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.590144.8863.9
k-NNNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.580746.751,113.4
GreedyNumber of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.57852.883,402.1
MLPBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.575242.3846.1
RandomBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.572642.2573.2
NeuralUCBBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.55841.1268.4
k-NNBackbone=Qwen3-0.6B, Number of seeds=5, Iterations=160, Warm-up phase=50-step2026.050.514636.2547.8