benchgap
Other · model

Glimpse benchmark scores

As of 2026-11-18, Glimpse (Other) has measured scores on 3 benchmarks and estimated scores on 23 more.

BenchmarkScoreSource
AA-MMMU-Pro85.0%estimated ± 1.9 pp, medium confidence
MMMU92.7%estimated ± 3.7 pp, low confidence
ParseBench-172.9%estimated ± 8.6 pp, low confidence
CharXiv92.9%measured
CharXiv w/o tools88.3%estimated ± 1.9 pp, medium confidence
AI2D_TEST93.5%estimated ± 0.9 pp, low confidence
CountBench100.0%estimated ± 2.2 pp, low confidence
DynaMath88.4%estimated ± 0.8 pp, low confidence
ERQA83.8%measured
MathVision94.5%estimated ± 3.3 pp, high confidence
MLVU (M-Avg)89.9%estimated ± 1.5 pp, high confidence
MMMU-Pro82.6%estimated ± 2.9 pp, high confidence
OmniDocBench 1.592.6%estimated ± 2.5 pp, low confidence
RealWorldQA88.3%estimated ± 1.5 pp, high confidence
ScreenSpot Pro89.3%estimated ± 6.8 pp, medium confidence
SimpleVQA77.8%estimated ± 9.2 pp, low confidence
V*99.8%estimated ± 11.2 pp, low confidence
Video-MME (w/o subtitle)93.4%estimated ± 10.9 pp, low confidence
Video-MME (with subtitle)91.8%estimated ± 2.9 pp, high confidence
VideoMMMU88.0%estimated ± 1.5 pp, medium confidence
We-Math96.6%estimated ± 2.1 pp, low confidence
MedXpertQA (MM)80.3%estimated ± 4.2 pp, low confidence
BabyVision w/ Python95.4%measured
BabyVision70.1%estimated ± 10.4 pp, low confidence
MMVU81.9%estimated ± 3.6 pp, medium confidence
MathVision w/ Python97.3%estimated ± 1.0 pp, low confidence