DeepMind has released a new image model called Gemini 3 Pro, which is part of their Nano Banana Pro toolkit. This model is designed for tasks such as image classification and object detection. According to the company, Gemini 3 Pro achieves state-of-the-art results on several benchmark datasets. The model's architecture is based on a combination of transformer and CNN components. --- Why it matters: This matters because it provides researchers with a powerful tool for improving image recognition capabilities in AI systems. By achieving top performance on various benchmarks, this model can help advance the field of computer vision. Source: https://deepmind.google/blog/build-with-nano-banana-pro-our-gemini-3-pro-image-model/