List your company

ResearchFormerly WizWand

MMDU

Benchmarks

TaskDatasetSOTA resultTrendResults

Multi-turn visual dialogue

MMDU 45K

3.79Accuracy

18

Multi-image Dialogue Understanding

MMDU

59.8Accuracy

18

Multimodal dialogue understanding

MMDU

0.703GPT-4o Score

10

Multi-turn Multi-image Dialog

MMDU

66.3Accuracy

4