Ranking according to https://paperswithcode.com/sota/image-classification-on-imagenet

At time of writing the best model is BASIC-L with 91.1%. Will a new model architecture emerge before the end of 2023 which beats this accuracy.

Must appear on the papers with code ranking.

Papers with Code - ImageNet Benchmark (Image Classification)

The current state-of-the-art on ImageNet is BASIC-L (Lion, fine-tuned). See a full comparison of 906 papers with code.