Skip to main content
ModelLLMsv2.0

InternVL 2

by Shanghai AI Laboratory · open-source · Last verified 2026-03-17

Shanghai AI Laboratory's scalable open-source vision-language model achieving GPT-4V-level performance on multimodal benchmarks. Features dynamic resolution support and progressive training from 1B to 108B parameter configurations.

https://github.com/OpenGVLab/InternVL
C
CBelow Average
Adoption: CQuality: B+Freshness: C+Citations: C+Engagement: F

Specifications

License
MIT
Pricing
open-source
Capabilities
image-understanding, visual-reasoning, dynamic-resolution, multi-image-understanding, ocr
Integrations
huggingface, vllm, transformers, lmdeploy
Use Cases
visual-qa, document-understanding, chart-analysis, multimodal-research
API Available
No
Parameters
108B
Context Window
32K tokens
Modalities
text, image
Training Cutoff
Mid 2024
Tags
multimodal, vision, open-source, shanghai-ai-lab, scalable
Added
2026-03-17
Completeness
90%

Index Score

46
Adoption
45
Quality
75
Freshness
58
Citations
52
Engagement
0

Explore the full AI ecosystem on Agents as a Service