X @Demis Hassabis

Model Capabilities - Gemini 3 Pro is an exceptional vision AI model [1] - The model achieves state-of-the-art (SOTA) performance across major vision and multimodal benchmarks [1] - It excels in understanding documents, screens, images, videos, and spatial contexts [1] Applications - The model is suitable for document understanding tasks [1] - It can be applied to screen understanding tasks [1] - It is also applicable to image and video understanding tasks [1] - The model can be used for spatial understanding tasks [1]