Multimodal AI Model Suite
10.0k 2026-04-18
OpenGVLab/InternVL
A pioneering open-source multimodal AI model family designed to serve as a high-performance alternative to commercial models like GPT-4o and GPT-5.
Core Features
State-of-the-art performance in multimodal, reasoning, text, and agentic tasks.
Comprehensive family of models (InternVL3.5, InternVL3, Mini-InternVL) with varying scales.
Open-sourced training code and datasets for research and development.
Enhanced versatility, reasoning capability, and inference efficiency.
Support for both GitHub and Hugging Face `transformers` model formats.
Detailed Introduction
InternVL is an acclaimed open-source multimodal AI model family, recognized with a CVPR 2024 Oral presentation, aiming to close the performance gap with leading commercial models such as GPT-4o and GPT-5. It offers a suite of models, including InternVL3.5, InternVL3, and Mini-InternVL, demonstrating state-of-the-art results across diverse multimodal, reasoning, and agentic tasks. The project emphasizes versatility, efficiency, and open-sourcing its training methodologies and datasets, fostering advanced research and application development in the multimodal AI domain.