MeanFlow has recently emerged as a powerful framework for few-step generative modeling trained from scratch, but its success is not yet fully understood. In this work, we show that the MeanFlow objective naturally decomposes into two parts: trajectory flow matching and trajectory consistency. Through gradient analysis, we find that these terms are strongly negatively correlated, causing optimization conflict and slow convergence. Motivated by these insights, we introduce α-Flow, a broad family of objectives that unifies trajectory flow matching, Shortcut Model, and MeanFlow under one formulation. By adopting a curriculum strategy that smoothly anneals from trajectory flow matching to MeanFlow, α-Flow disentangles the conflicting objectives, and achieves better convergence. When trained from scratch on class-conditional ImageNet-1K 256x256 with vanilla DiT backbones, α-Flow consistently outperforms MeanFlow across scales and settings. Our largest α-Flow-XL/2+ model achieves new state-of-the-art results using vanilla DiT backbones, with FID scores of 2.58 (1-NFE) and 2.15 (2-NFE).
@article{arxiv.2510.20771,
title = {AlphaFlow: Understanding and Improving MeanFlow Models},
author = {Huijie Zhang and Aliaksandr Siarohin and Willi Menapace and Michael Vasilkovsky and Sergey Tulyakov and Qing Qu and Ivan Skorokhodov},
journal= {arXiv preprint arXiv:2510.20771},
year = {2025}
}