ProxyThinker: Test-Time Guidance through Small Visual Reasoners

ICLR 2026 Conference Submission15310 Authors

Keywords: decoding-time algorithms, visual reasoning

TL;DR: ProxyThinker enables large vision-language models to inherit slow-thinking visual reasoning skills from smaller RFT models at inference time.

Abstract: Recent advancements in reinforcement learning with verifiable rewards have pushed the boundaries of the visual reasoning capabilities in large vision-language models (LVLMs). However, training LVLMs with reinforcement fine-tuning (RFT) is computationally expensive, posing a significant challenge to scaling model size. In this work, we propose ProxyThinker, an inference-time technique that enables large models to inherit the visual reasoning capabilities from small, slow-thinking visual reasoners without any training. By subtracting the output distributions of base models from those of RFT reasoners, ProxyThinker modifies the decoding dynamics and successfully elicits the slow-thinking reasoning demonstrated by the emerged sophisticated behaviors such as self-verification and self-correction. ProxyThinker consistently boosts performance on challenging visual benchmarks on spatial, mathematical, and multidisciplinary reasoning, enabling untuned base models to compete with the performance of their full-scale RFT counterparts. Furthermore, our implementation efficiently coordinates multiple language models with parallelism techniques and achieves faster inference compared to previous decoding-time methods, paving the way for the practical deployment of ProxyThinker. Code is available at https://anonymous.4open.science/r/ProxyThinker-FAAF.

Primary Area: applications to computer vision, audio, language, and other modalities

Submission Number: 15310