PhiZero: A World Model Built Around Physical Language
For researchers in video generation and world modeling, PhiZero offers a new paradigm that improves physical coherence by leveraging explicit language-like reasoning, potentially addressing a known bottleneck in existing pixel-based models.
PhiZero introduces a physical world model that uses a compact discrete representation called physical language to explicitly reason about world-state transitions before rendering videos, rather than predicting pixels directly. Experiments across generation and understanding benchmarks demonstrate its ability to model physically coherent world evolution, with potential for interactive simulation and zero-shot motion transfer.
We introduce PhiZero, a physical world model built around physical language, a compact discrete representation of world-state transitions. Existing physical world models typically predict future videos directly in pixel space, leaving the underlying world dynamics implicit within high-dimensional visual predictors. Motivated by humans' ability to abstract predictive structure from visual experience and organize it in natural language for explicit reasoning, we learn physical language from in-the-wild videos through self-supervision and use it to explicitly reason about how the physical world evolves. Accordingly, PhiZero adopts a reason-then-render paradigm: it first infers future world evolution as a physical-language sequence and then renders the inferred transitions into videos. Extensive experiments across generation and understanding benchmarks validate the ability of PhiZero to model physically coherent world evolution. We further show its potential for realistic and interactive world modeling, fine-grained action-conditioned simulation, and zero-shot motion transfer.