LFM2.5-VL-3B

LFM2.5-VL-3B places in the top three on POPE.

Liquid AI Released 12 Aug 2026 Family LFM2.5 The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Liquid AI Aug 2026

Open weights Small

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
POPE	88.7%	#2 of 7	Model card
MMMB	83%	#1 of 2	Model card
MM-IFEval	60.6%	#2 of 3	Model card
IFEval	82.3%	#90 of 126	Model card
Multi-IF	59.4%	#23 of 28	Model card
IFBench	25.8%	#69 of 70	Model card
HallusionBench	47.2%	#23 of 23	Model card
SimpleVQA	35.4%	#23 of 23	Model card

Benchmark	Score	Rank	Source
ScreenSpot-v2	80.7%	#13 of 14	Model card
BFCL V4	32.5%	#24 of 24	Model card
ToolSandbox	59.5%	#5 of 5	Model card

Benchmark	Score	Rank	Source
TextVQA	84.3%	#4 of 17	Model card
SEED-Bench	77.7%	#1 of 3	Model card
OCRBench v2	42.6%	#35 of 65	Official leaderboard
BLINK	61.5%	#17 of 28	Model card
RealWorldQA	73.1%	#21 of 30	Model card
MathVista	68.5%	#45 of 64	Model card
DocVQA	91.1%	#25 of 33	Model card
ChartQA	81.3%	#29 of 37	Model card
RefCOCO / RefCOCO+ / RefCOCOg	87.9%	#15 of 19	Model card
CountBench	87.3%	#16 of 20	Model card
InfographicVQA	70.2%	#18 of 22	Model card
MuirBench	58.3%	#10 of 12	Model card
OCRBench	84.2%	#24 of 28	Model card
MMStar	63.3%	#28 of 30	Model card
MMBench-EN v1.1	81%	#16 of 17	Model card
MMMU	48.4%	#92 of 97	Model card
MMMU-Pro	30.5%	#71 of 72	Model card
LogicVista	37.4%	#12 of 12	Model card

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