Canvas-of-Thought: Grounding Reasoning via Mutable Structured States

TL;DR

Canvas-of-Thought introduces an external visual reasoning substrate using HTML Canvas, enabling explicit state management and visual feedback, which enhances multimodal reasoning in large language models for complex tasks.

Contribution

The paper proposes Canvas-CoT, a novel framework that uses a HTML Canvas for explicit, in-place state updates and visual critique, improving reasoning efficiency and accuracy.

Findings

Outperforms existing baselines on VCode, RBench-V, and MathVista datasets.

Enables in-place state revisions without context disruption.

Provides explicit visual feedback to improve reasoning in high-dimensional domains.

Abstract

While Chain-of-Thought (CoT) prompting has significantly advanced the reasoning capabilities of Multimodal Large Language Models (MLLMs), relying solely on linear text sequences remains a bottleneck for complex tasks. We observe that even when auxiliary visual elements are interleaved, they are often treated as static snapshots within a one-dimensional, unstructured reasoning chain. We argue that such approaches treat reasoning history as an immutable stream: correcting a local error necessitates either generating verbose downstream corrections or regenerating the entire context. This forces the model to implicitly maintain and track state updates, significantly increasing token consumption and cognitive load. This limitation is particularly acute in high-dimensional domains, such as geometry and SVG design, where the textual expression of CoT lacks explicit visual guidance, further constraining the model's reasoning precision. To bridge this gap, we introduce \textbf{Canvas-of-Thought (Canvas-CoT)}. By leveraging a HTML Canvas as an external reasoning substrate, Canvas-CoT empowers the model to perform atomic, DOM-based CRUD operations. This architecture enables in-place state revisions without disrupting the surrounding context, allowing the model to explicitly maintain the "ground truth". Furthermore, we integrate a rendering-based critique loop that serves as a hard constraint validator, providing explicit visual feedback to resolve complex tasks that are difficult to articulate through text alone. Extensive experiments on VCode, RBench-V, and MathVista demonstrate that Canvas-CoT significantly outperforms existing baselines, establishing a new paradigm for context-efficient multimodal reasoning.