Spectral-Progressive Thought Flow for Lightweight Multimodal Reasoning

Multimodal reasoning often relies on long chains of intermediate textual and visual thoughts, where accumulating visual tokens and dense cross-modal attention incur substantial computation and memory overhead. To address this challenge, we propose Spectral-Progressive Thought Flow (*SpecFlow*), a *novel* lightweight multimodal reasoning framework that represents intermediate visual thoughts in a fixed-size discrete cosine space. By exploiting strong energy compaction, *SpecFlow* preserves global layout and relational structure while introducing high-frequency details only when increased spatial precision is required. To align visual state evolution with linguistic intent, classifier-free guidance enables autoregressive textual thoughts to steer flow-based updates of the visual workspace without expanding the context. As a result,*SpecFlow* maintains a bounded visual workspace whose updates depend only on the current visual state and accumulated textual trace, enabling long-horizon inference with stable latency and memory usage independent of reasoning depth. Empirical results show that *SpecFlow* achieves competitive or superior reasoning performance while reducing computation and memory costs by up to *$2.1\times$*.

AI systems often need to reason over both images and text, such as finding paths, tracking objects, or understanding spatial relationships, but existing methods can become slow and memory-heavy because they store many intermediate visual steps. This paper introduces SpecFlow, an efficient approach that keeps only a compact visual workspace and updates it step by step, focusing first on the scene’s global structure and adding finer details only when needed. Guided by the text question, SpecFlow supports strong visual-spatial reasoning while using less computation and memory, making long-step multimodal reasoning more practical.