The Year in AI Papers
The year/Independent research

Paper 2603.00141

From Scale to Speed: Adaptive Test-Time Scaling for Image Editing

Xiangyan Qu, Zhenlong Yuan, Jing Tang, Rui Chen, Datao Tang, Meng Yu, Lei Sun, Yancheng Bai, Xiangxiang Chu, Gaopeng Gou, Gang Xiong, Yujun Cai

Published
Feb 2026
Research lab
Independent
Citations
6
GitHub
Not linked

01 In brief

Summary

The paper introduces ADE-CoT, a test-time scaling framework for image editing that improves efficiency and performance.

It addresses three issues with applying Image-CoT to editing: inefficient fixed sampling budgets, unreliable early-stage verification using general MLLM scores, and redundant results from large-scale sampling.

ADE-CoT uses difficulty-aware resource allocation to adjust sampling budgets based on edit difficulty, edit-specific verification (region localization and caption consistency) for accurate early pruning, and depth-first opportunistic stopping guided by an instance-specific verifier to terminate when intent-aligned results are found.

Experiments on three editing models (Step1X-Edit, BAGEL, FLUX.1 Kontext) across three benchmarks show ADE-CoT achieves better performance with over 2x speedup compared to Best-of-N under comparable budgets.

The framework also improves reasoning efficiency (η) and outcome efficiency (ξ) metrics, indicating better performance-efficiency trade-offs and reduced redundancy.

Ablations confirm the contribution of each component, and the method is robust to different MLLM verifiers, with stronger MLLMs yielding higher gains.

The work highlights the potential of adaptive test-time scaling for goal-directed generation tasks beyond text-to-image.

02 From the paper

Abstract

Image Chain-of-Thought (Image-CoT) is a test-time scaling paradigm that improves image generation by extending inference time. Most Image-CoT methods focus on text-to-image (T2I) generation. Unlike T2I generation, image editing is goal-directed: the solution space is constrained by the source image and instruction. This mismatch causes three challenges when applying Image-CoT to editing: inefficient resource allocation with fixed sampling budgets, unreliable early-stage verification using general MLLM scores, and redundant edited results from large-scale sampling. To address this, we propose ADaptive Edit-CoT (ADE-CoT), an on-demand test-time scaling framework to enhance editing efficiency and performance. It incorporates three key strategies: (1) a difficulty-aware resource allocation that assigns dynamic budgets based on estimated edit difficulty; (2) edit-specific verification in early pruning that uses region localization and caption consistency to select promising candidates; and (3) depth-first opportunistic stopping, guided by an instance-specific verifier, that terminates when intent-aligned results are found. Extensive experiments on three SOTA editing models (Step1X-Edit, BAGEL, FLUX.1 Kontext) across three benchmarks show that ADE-CoT achieves superior performance-efficiency trade-offs. With comparable sampling budgets, ADE-CoT obtains better performance with more than 2x speedup over Best-of-N.