Gemma 3n is a multimodal AI model. It natively supports image, audio, video, and text inputs. However, it can only generate text outputs. It is also a multilingual model and supports 140 languages for ...