WorldCast: Exploring Intelligent Visual Generation in World State Prediction
WorldCast ranks 12 models on visual world-state prediction, with GPT-Image-2.5 scoring 72.8.
WorldCast is a benchmark of 320 manually designed cases across eight categories for multi-step visual world-state prediction from multimodal observations. Across 12 models, GPT-Image-2.5 System scores 72.8, Nano Banana Pro scores 42.8, and the best open-source model scores 20.1. The authors report early frontier capability but large deficits on complex context, entangled references, and reasoning-heavy tasks, and note that strong generators draw on their underlying language models.
- WorldCast has 320 cases across eight categories.
- GPT-Image-2.5 System scores 72.8.
- Nano Banana Pro scores 42.8; best open model scores 20.1.
- Large gaps remain on complex context and reasoning-heavy tasks.
Full article252 words · extracted from huggingface.co · click to collapse
Humans interpret multimodal observations of the world and forecast how it will evolve. Recent multimodal generative systems have demonstrated similar capabilities: they can integrate multimodal context and synthesize visual outputs that depict changes in the world. To characterize this emerging trajectory, we organize model capabilities into a four-level hierarchy and discover that higher-level tasks remain relatively underexplored despite model progress. To bridge this gap, we explore model performance on multi-step world state prediction, a high-level task that requires models to forecast successive future world states from multimodal observations and diverse references. Such a task requires models to possess entangled understanding and generation capabilities, as well as extensive world knowledge to translate human action sequences, physical transformations, and abstract visual reasoning into coherent visual outputs. We introduce WorldCast, a benchmark comprising 320 manually designed cases across eight categories, and evaluate 12 representative models. The conducted experiments yield three key insights: (a) advances in frontier models are rapidly pushing the boundaries of intelligent world modeling, with the GPT-Image-2.5 System scoring 72.8 and demonstrating early-stage visual state prediction capabilities on real-world tasks; (b) WorldCast reveals substantial gaps between models: even Nano Banana Pro reaches only 42.8 and the best open-source model scores 20.1, with deficits in handling complex context, entangled references, and reasoning-heavy tasks; (c) strong generative models effectively draw on the capabilities of their underlying language models. These findings highlight the need for multimodal generative models that are capable of dynamic planning and reasoning while also being deeply unified to faithfully simulate real-world states.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2610.20731