Which is better, GPT-4o or InternVL-2-4B?

Refresh Models —Deals —Regulatory —Sources —

Questions this page answers

Which is better, GPT-4o or InternVL-2-4B?: Across 3 shared benchmarks, GPT-4o scores higher on 3 and InternVL-2-4B on 0. The widest gap is MMMU (val) (Pass@1), where GPT-4o scores 69.1 against 44.2.

GPT-4o vs InternVL-2-4B

Across 3 shared benchmarks, GPT-4o scores higher on 3 and InternVL-2-4B on 0. The widest gap is MMMU (val) (Pass@1), where GPT-4o scores 69.1 against 44.2.

OpenAIvsOpenGVLab3 shared benchmarks3–0 head-to-head

Benchmark	GPT-4o	InternVL-2-4B
AI2D	94.2	77.3
MathVista	63.8	53.7
MMMU (val) (Pass@1)	69.1	44.2

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.