The Shanghai AI Laboratory has recently unveiled and open-sourced its latest version, Scholar·Universe 3.0 (InternVL3), a universal multimodal large model. This model employs groundbreaking pre-training and fine-tuning methodologies to holistically bolster its multimodal foundational capabilities. In numerous tests, Scholar·Universe 3.0 has emerged as the top performer among open-source models, markedly enhancing functionalities such as GUI agents and drawing comprehension.
