InternVL3 78B

The InternVL3 collection of models are advanced multimodal large language models that combine superior vision and language understanding capabilities. Built with a ViT-MLP-LLM architecture, these models excel at multimodal reasoning, document analysis, video understanding, and complex visual tasks while supporting dynamic resolution processing and extended context understanding.