VECTOR WIREAI INTELLIGENCE
UTC
Refresh Models Deals Regulatory Sources

InternVL-3-78B vs LFM2.5-VL-3B

9 SHARED BENCHMARKS

Across 9 shared benchmarks, InternVL-3-78B scores higher on 5 and LFM2.5-VL-3B on 4. The widest gap is MMMU (val) (Pass@1), where InternVL-3-78B scores 60.8 against 48.4.

OPENGVLABVSLIQUID AI9 SHARED54 HEAD-TO-HEAD
BLINK51.961.5
DocVQA-val83.891.1
MathVista70.168.5
MMStar66.163.3
POPE88.988.7
RealWorldQA74.373.1
TextVQA-val83.584.3

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.