Open MLLMs excelling in vision, reasoning & long context
Open MLLM family (1B-78B) from OpenGVLab. Excels at vision, reasoning, long context & agents via native multimodal pre-training. Outperforms base LLMs on text tasks.
