Inference
EN

MMLongBench-Doc

A long-document multimodal benchmark for grounded reasoning over extended document contexts.

ModelsModels with a result.6
Top resultBest result on this benchmark.—
Top-3 spreadPoints from first to third.—
YearYear of release.—

Results

6 results
#ModelResultThe score from the source.
—Nemotron 3.5 Super VL 120B A12B62.8%
—Nemotron 3.5 Super VL 120B A12B NVFP462.8%
—Qwen3.5 397B61.5%
—Gemini 3 Pro60.5%
—Kimi K2.558.5%
—Nemotron 3 Nano Omni 30B A3B57.5%