InternVL2_5-2B-MPO is part of a series of multimodal large language models that demonstrate exceptional overall performance. This series is built on InternVL2.5 and mixed preference optimization. It integrates InternViT, which has undergone incremental pre-training, with various pre-trained large language models, including InternLM 2.5 and Qwen 2.5, utilizing randomly initialized MLP projectors. The model excels in multimodal tasks, capable of processing diverse data types, including images and text, and is suitable for scenarios requiring the comprehension and generation of multimodal content.