Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced reasoning capabilities in Large Language Models. However, adapting RLVR to multimodal domains suffers from a critical \textit{perception-reasoning decoupling}. Existing paradigms, driven by text-centric outcome rewards, reasoning in language medium, inadvertently encourage models to bypass visual perception. We empirically validate this through blind experiments: state-of-the-art policies maintain or surprisingly improve performance even when visual inputs are entirely removed. This reveals that these models degenerate into \textit{blind reasoners}, exploiting linguistic priors to generate plausible answers instead of attending to visual evidence. In response, we propose \textbf{Thinking with Deltas}, a framework driven by a \textbf{Differential Visual Reasoning Policy (DVRP)}. DVRP introduces intrinsic supervision via visual triplets, comprising original, masked, and perturbed inputs. It optimizes the model to maximize reasoning divergence from masked inputs (enforcing \textit{visual sensitivity}) while minimizing divergence from perturbed inputs (ensuring \textit{visual robustness}). By aligning reasoning variations strictly with the \textit{Delta} of visual information, DVRP inherently bolsters visual understanding capabilities and significantly outperforms state-of-the-art methods on both general and medical benchmarks, without requiring external annotations or auxiliary tools.

翻译：可验证奖励强化学习（RLVR）显著提升了大型语言模型的推理能力。然而，将RLVR应用于多模态领域时，存在关键的“感知-推理解耦”问题。现有范式以文本为中心的结果奖励为驱动，在语言媒介中进行推理，无意中鼓励模型绕过视觉感知。我们通过盲实验对此进行了实证验证：即使视觉输入被完全移除，最先进的策略仍能保持甚至意外地提升性能。这表明这些模型退化为“盲推理器”，利用语言先验生成看似合理的答案，而非关注视觉证据。为此，我们提出**基于差分思维**的框架，该框架由**差分视觉推理策略（DVRP）**驱动。DVRP通过视觉三元组（包含原始、掩码和扰动输入）引入内在监督。它优化模型以最大化与掩码输入之间的推理差异（强制“视觉敏感性”），同时最小化与扰动输入之间的推理差异（确保“视觉鲁棒性”）。通过将推理变化严格与视觉信息的“差分”对齐，DVRP本质上增强了视觉理解能力，并在通用和医学基准测试中显著优于现有最先进方法，且无需外部标注或辅助工具。