cs.AINov 19, 2026

Keystone: Judging Previews at Key Steps for Test-Time Search in Diffusion Models

Authors: Jiajun Fan, Chaoran Cheng, Ziwen Wang, Changyi Yang, Bangji Yang, Qinglong Zheng, Ge Liu

Organizations: UIUC · NYU · CMU

Abstract

Text-to-image diffusion and flow models yield a preview, their clean-image estimate, at every denoising step. When they miscount, drop or misplace objects, test-time search is in a bind: its verifiers are reward models built for finished images, so judging early misreads blurry previews, and judging late, as best-of-N does, pays to fully denoise doomed candidates. We find that the image is not the problem: the outcome is settled as denoising starts, by the seed or by a stochastic sampler's first steps, and already shows in early previews; what fails on them is the judge, as reward models read them near chance on several kinds of prompts. We introduce Keystone, a training-free judge of previews that compiles a prompt into typed questions about its objects, counts, colours and positions and has an off-the-shelf multimodal decision model answer them. Asked this way, the judge reads counts and positions from early previews better than reward models read them from finished images. Searching at these key steps on GenEval with SD3.5-Medium, Keystone's adaptive search and pruning significantly outperform best-of-8 with every other verifier at under half its generator compute, and as a drop-in verifier there, the Keystone score significantly improves best-of-N, progressive seed pruning and FK steering; without retuning, both searches stay significantly above the strongest other verifier's best-of-N on average over four base models and three benchmarks. Keystone thus moves test-time judgment to the key steps where the outcome already shows.

Figures & tables

Appendix figures & tables24 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. PreviewDiff: Multimodal Critic-Guided Search over Diffusion Latents

    Sep 28, 2026Vighnesh Subramaniam, Boris Katz, Brian Cheung +3Diffusion Model GuidanceT2I Generation

  2. Toward Early Quality Assessment of Text-to-Image Diffusion Models

    Mar 3, 2026Huanlei Guo, Hongxin Wei, Bingyi JingT2I Generation EvaluationT2I Diffusion Models

  3. Inference-Time Scaling of Diffusion Models via Progressive Seed Pruning

    Jul 23, 2026Rogerio Guimaraes, Pietro PeronaDiffusion Model SamplingDiffusion Sampling