GLM-4.5V

GLM-4.5V places in the top three on 2 benchmarks.

Z.ai Released 11 Aug 2025 Family GLM-V The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Z.ai Aug 2025

Open weights Mid-size

No Epoch Capabilities Index score yet

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score	Rank	Source
HallusionBench	65.4%	#8 of 23	Model card
MMBench-CN v1.1	88.3%	#2 of 5	Model card
LMArena (Elo)	1,352	#174 of 333	Official leaderboard

Benchmark	Score		Rank	Source
HiPhO	35.3%		#25 of 35	Official leaderboard

Benchmark	Score		Rank	Source
Flame-React-Eval	82.5%		#2 of 3	Model card

Benchmark	Score	Rank	Source
WebVoyager	84.4%	#6 of 11	Model card
SnakeBench	21	#116 of 183	Official leaderboard
OSWorld	35.8%	#9 of 12	Model card
AndroidWorld	57%	#15 of 18	Model card

Benchmark	Score	Rank	Source
MMMU	75.4%	#20 of 136	Official leaderboard
MathVista	84.6%	#16 of 64	Model card
CVBench	87.3%	#3 of 11	Model card
Ref-L4	89.5%	#1 of 3	Model card
Math-VR	49.6%	#10 of 26	Official leaderboard
BLINK	65.3%	#12 of 28	Model card
RefCOCO / RefCOCO+ / RefCOCOg	91.3%	#9 of 19	Model card
MuirBench	75.3%	#6 of 12	Model card
GeoBench	79.7%	#1 of 2	Model card
ZeroBench	23.4%	#13 of 25	Model card
MMStar	75.3%	#16 of 30	Model card
OCRBench	86.5%	#15 of 28	Model card
Video-MME	80.7%	#23 of 42	Model card
VideoMME v2 (with subtitles)	21.8%	#17 of 31	Official leaderboard
AI2D	88.1%	#25 of 44	Model card
MathVerse	72.1%	#7 of 12	Model card
MMBench-EN v1.1	88.2%	#10 of 17	Model card
MATH-Vision (MathVision)	65.6%	#27 of 42	Model card
LogicVista	62.4%	#8 of 12	Model card
All-Angles Bench	56.9%	#4 of 6	Model card
WebQuest	68.8%	#2 of 3	Model card
MMMU-Pro	65.2%	#49 of 72	Model card
MMVU	68.7%	#14 of 20	Model card
TreeBench	50.1%	#5 of 7	Model card
We-Math	68.8%	#8 of 11	Model card
Vision Arena	1,153	#99 of 136	Official leaderboard
ChartQA Pro	64%	#6 of 8	Model card
DynaMath	53.9%	#13 of 17	Model card
MMLongBench-Doc	44.7%	#10 of 13	Model card
MVBench	73%	#12 of 15	Model card
ChartMuseum	55.3%	#4 of 5	Model card
OmniSpatial	51%	#4 of 5	Model card
Video-MMMU	72.4%	#25 of 30	Model card
ERQA	50%	#22 of 26	Model card
LVBench	53.8%	#22 of 25	Model card
MotionBench	62.4%	#8 of 9	Model card
Math Kangaroo 2025 (MathArena)	67.6%	#21 of 22	Official leaderboard

The benchmarks the lab chose to headline, with the scores it gave.

Every result in the longer documents the lab published with it.