MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis

MiMo-v2.6-Pro is a computer vision model that claims to outperform SOTA (state-of-the-art) models on various benchmarks. It has 2.3 billion parameters and is based on the Swin Transformer architecture. The model requires 8 NVIDIA V100 GPUs for training, with a total of 256 GB of VRAM. It was trained on a dataset of 1.5 million images from the ImageNet dataset. MiMo-v2.6-Pro achieved top-1 accuracy of 88.3% and top-5 accuracy of 97.8% on ImageNet, outperforming SOTA models by a significant margin.