Which is better, GPT-4o mini or InternVL-2-8B?

Refresh Models —Deals —Regulatory —Sources —

Questions this page answers

Which is better, GPT-4o mini or InternVL-2-8B?: Across 10 shared benchmarks, GPT-4o mini scores higher on 5 and InternVL-2-8B on 5. The widest gap is InterGPS (test), where InternVL-2-8B scores 53.2 against 39.9.

GPT-4o mini vs InternVL-2-8B

Across 10 shared benchmarks, GPT-4o mini scores higher on 5 and InternVL-2-8B on 5. The widest gap is InterGPS (test), where InternVL-2-8B scores 53.2 against 39.9.

OpenAIvsOpenGVLab10 shared benchmarks5–5 head-to-head

Benchmark	GPT-4o mini	InternVL-2-8B
AI2D	75.2	81.4
BLINK	51.9	45.4
InterGPS (test)	39.9	53.2
MathVista	56.7	51.1
MMBench (dev-en)	83.8	87
MMMU (val) (Pass@1)	52.1	46.3
POPE (test)	83.6	84.2
ScienceQA (img-test)	84	95.9
TextVQA (val)	70.9	68.8
Video-MME Overall	61.2	52.6

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.