List your company

ResearchFormerly WizWand

MMLongBench-Doc

Benchmarks

TaskDatasetSOTA resultTrendResults

Multimodal Long-document Understanding

MMLongBench-Doc 1.0 (test)

46.7Reports Score

12

MMLongBench-Doc 2K Packed

55.66Accuracy (Residential)

11

Question Answering

MMLongBench-Doc Multi-page (MP)

24.18EM

10

Document Retrieval

MMLongBench-Doc Multi-page (MP)

61.38Recall

10

Question Answering

MMLongBench-Doc Single-page

29.34EM

10

Document Retrieval

MMLongBench-Doc Single-page (SP)

75.77Top-1 Accuracy

10

Long-context document understanding

MMLongBench-Doc 32K context slice

45MMLongBench-Doc (32K) Score

4

Long-context document understanding

MMLongBench-Doc 16K context slice

48.57Score

4