MMGR: Multi-Modal Generative Reasoning
# MMGR: Multi-Modal Generative Reasoning
**Auteur:** Zefan Cai
**Date de publication:** 16 décembre 2025
**Source:** [Hugging Face (paper)](https://huggingface.co/papers/2512.14691)
**Score de pertinence:** 89/100
—
## À propos
Video foundation models generate visually realistic and temporally coherent content, but their reliability as world simulators depends on whether they capture physical, logical, and spatial constraints. Existing metrics such as Frechet Video Distance (FVD) emphasize perceptual quality and overlook reasoning failures, including violations of causality, physics, and global consistency. We introduce MMGR (Multi-Modal Generative Reasoning Evaluation and Benchmark), a principled evaluation framework based on five reasoning abilities: Physical, Logical, 3D Spatial, 2D Spatial, and Temporal. MMGR evaluates generative reasoning across three domains: Abstract Reasoning (ARC-AGI, Sudoku), Embodied Navigation (real-world 3D navigation and localization), and Physical Commonsense (sports and compositional interactions). MMGR applies fine-grained metrics that require holistic correctness across both video and image generation. We benchmark leading video models (Veo-3, Sora-2, Wan-2.2) and image models (Nano-banana, Nano-banana Pro, GPT-4o-image, Qwen-image), revealing strong performance gaps across domains. Models show moderate success on Physical Commonsense tasks but perform poorly on Abstract Reasoning (below 10 percent accuracy on ARC-AGI) and struggle with long-horizon spatial planning in embodied settings. Our analysis highlights key limitations in current models, including overreliance on perceptual data, weak global state consistency, and objectives that reward visual plausibility over causal correctness. MMGR offers a unified diagnostic benchmark and a path toward reasoning-aware generative world models.
## Mots-clés détectés
– benchmark
– GPT
– AGI
– reasoning
– performance
– accuracy
– image generation
## En savoir plus
Pour plus de détails sur cette découverte, consultez l’article original : [MMGR: Multi-Modal Generative Reasoning](https://huggingface.co/papers/2512.14691)
—
*Cet article a été généré automatiquement par le système de veille IA de Sophia.*