List your company

Multi-task Evaluation on Aggregate All Tasks

74.6ScoreEvoRoute
51.51257.50663.569.494Dec 21, 2022Jun 24, 2023Dec 26, 2023Jun 29, 2024Dec 31, 2024Jul 4, 2025Jan 6, 2026

Evaluation results

MethodLinks
EvoRouteSetting=Ours2026.0174.6138.492.23
Gemini-2.5-proSetting=Manual2026.0174.27521.61123
MasRouterSetting=Routing2026.0171.7176.9897.2
GraphRouterSetting=Routing2026.0167.13208.8100.8
VMSSTInitialization=Pretrained, Number of Layers=242022.1265.9——
PromptLLMSetting=Routing2026.0165.76200.73100.43
BITRANSLATIONInitialization=Pretrained, Number of Layers=242022.1264.9——
VMSST CONTRASTIVEInitialization=Pretrained, Number of Layers=242022.1264.4——
CONTRASTIVEInitialization=Pretrained, Number of Layers=242022.1264.2——
GPT-4oSetting=Manual2026.0160.94581.08113.92
VMSSTInitialization=Random, Number of Layers=242022.1258.1——
GPT-4.1Setting=Manual2026.0157.23352.63112.73
BITRANSLATIONInitialization=Random, Number of Layers=242022.1255.9——
VMSSTInitialization=Random, Number of Layers=62022.1255.6——
VMSST CONTRASTIVEInitialization=Random, Number of Layers=242022.1255——
CONTRASTIVEInitialization=Random, Number of Layers=62022.1253.7——
VMSST CONTRASTIVEInitialization=Random, Number of Layers=62022.1253.7——
Qwen3-14bSetting=Manual2026.0153.6249.878.05
BITRANSLATIONInitialization=Random, Number of Layers=62022.1252.5——
CONTRASTIVEInitialization=Random, Number of Layers=242022.1252.4——