benchgap
Zhipu (GLM) · model

GLM-4.6V-Flash benchmark scores

As of 2026-11-19, GLM-4.6V-Flash (Zhipu (GLM)) has measured scores on 4 benchmarks and estimated scores on 20 more.

BenchmarkScoreSource
AA-MMMU-Pro60.0%estimated ± 1.5 pp, high confidence
MMMU71.1%measured
ParseBench-140.9%estimated ± 8.4 pp, medium confidence
ParseBench-246.9%estimated ± 5.0 pp, medium confidence
CharXiv54.5%estimated ± 5.1 pp, low confidence
CharXiv w/o tools71.8%estimated ± 2.1 pp, high confidence
AI2D_TEST88.8%estimated ± 2.0 pp, high confidence
CountBench93.6%estimated ± 3.0 pp, high confidence
DynaMath81.4%estimated ± 5.3 pp, low confidence
ERQA47.1%estimated ± 5.2 pp, low confidence
MathVision67.6%estimated ± 2.9 pp, high confidence
MLVU (M-Avg)78.1%estimated ± 2.2 pp, high confidence
MMMU-Pro60.6%measured
RealWorldQA80.0%estimated ± 2.8 pp, high confidence
ScreenSpot Pro55.3%estimated ± 7.9 pp, low confidence
SimpleVQA53.0%estimated ± 7.1 pp, medium confidence
Video-MME (w/o subtitle)72.6%estimated ± 1.2 pp, high confidence
Video-MME (with subtitle)77.9%estimated ± 3.0 pp, high confidence
VideoMMMU70.1%measured
We-Math75.4%estimated ± 7.2 pp, low confidence
MedXpertQA (MM)28.4%estimated ± 2.4 pp, medium confidence
MMVU65.6%estimated ± 2.7 pp, medium confidence
OCRBench V263.5%measured
MMMU-Pro w/ Python68.8%estimated ± 2.6 pp, medium confidence