The rapid advancement of native multi-modal models and omni-models, exemplified by GPT-4o, Gemini and o3 with their capability to process and generate content across modalities such as text and images, marks a significant milestone in the evolution of intelligence. Systematic evaluation of their multi-modal output capabilities in visual thinking process (a.k.a., multi-modal chain of thought, M-CoT) becomes critically important. However, existing benchmarks for evaluating multi-modal models primarily focus on assessing multi-modal inputs and text-only reasoning process while neglecting the importance of reasoning through multi-modal outputs. In this paper, we present a benchmark, dubbed as RBench-V, designed to assess models’ vision-indispensable reasoning. To conduct RBench-V, we carefully hand-pick 803 questions covering math, physics, counting and games. Unlike problems in previous benchmarks, which typically specify certain input modalities, RBench-V presents problems centered on multi-modal outputs, which require image manipulation, such as generating novel images and constructing auxiliary lines to support reasoning process. We evaluate numerous open- and closed-source models on RBench-V, including o3, Gemini 2.5 pro, Qwen2.5-VL, etc. Even the best-performing model, o3, achieves only 25.8% accuracy on RBench-V, far below the human score of 82.3%, which shows current models struggle to leverage multi-modal reasoning. Data and code are available at https://evalmodels.github.io/rbenchv.

We present a novel multigrid solver framework that significantly advances the efficiency of physical simulation for unstructured meshes. While multi-grid methods theoretically offer linear scaling, their practical implementation for deformable body simulations faces substantial challenges, particularly on GPUs. Our framework achieves up to 6.9× speedup over traditional methods through an innovative combination of matrix-free vertex block Jacobi smoothing with a Full Approximation Scheme (FAS), enabling both piecewise constant and linear Galerkin formulations without the computational burden of dense coarse matrices. Our approach demonstrates superior performance across varying mesh resolutions and material stiffness values, maintaining consistent convergence even under extreme deformations and challenging initial configurations. Comprehensive evaluations against state-of-the-art methods confirm our approach achieves lower simulation error with reduced computational cost, enabling simulation of tetrahedral meshes with over one million vertices at approximately one frame per second on modern GPUs.