While modern video generative models achieve remarkable visual fidelity, they remain noticeably brittle when generating physically plausible dynamics and complex interactions. Targeted post-training is crucial, yet supervised fine-tuning suffers from offline state mismatch, and reinforcement learning is hindered by sparse, ill-defined physical reward functions. On-Policy (Self-)Distillation (OPSD) reconciles on-policy rollouts with dense corrective feedback, shifting the core challenge to designing effective privileged context. However, conventional modalities fall short: text prompts are inherently declarative, whereas real reference videos heavily entangle physical laws with visual appearance without parametric control. To this end, we propose PhysOPSD, introducing executable code as privileged physical context for video generation. Unlike static visual observations, executable code instantiates an editable physical world with intrinsic abstraction and controllability. Powered by PhysOPSD-Engine, we translate real-world physical scenarios into verifiable simulation programs. PhysOPSD then incorporates (i) dual-level physical privilege to bridge the information gap via an intermediate bottleneck, alongside (ii) controlled physical variations under shared rollout noise to expose causal dynamic shifts. Extensive experiments demonstrate that PhysOPSD substantially improves physical plausibility across diverse backbones, outperforming SFT and RL baselines with only 3K samples, while preserving general video generation quality.
Figures & tables
Figure 2: Background overview. ( a ) OPSD shifts physical post-training from reward engineering to privileged context design; ( b ) Executable code explicitly parameterizes physical processes, uniquely unlocking abstraction and controllability ; ( c ) Using executable code as privileged context , PhysOPSD internalizes physical dynamics into the student for simulation-free video generation.
Figure 3: Overview of our proposed PhysOPSD .
Figure 4: Case study on physical context and self-teacher generation. From a swing-set reference, PhysOPSD constructs dual-level drafts ( ddense,dabs ) across paired variants varying only in initial angular velocity ( 0.12 vs . 1.7 rad/s ). Conditioned on these drafts, the self-teacher faithfully synthesizes the corresponding motion dynamics. More case studies are in Figures 11 , 12 and 13 .
Method
Size
Post-Train
PhyGround [ 24 ]
VideoPhy-2 [ 3 ]
VBench [ 11 ]
Data
General
Physics
Overall
SA
PC
SA ≥4
PC ≥4
All
Total
Quality
Semantic
Veo 3.1 [ 44 ]
-
71.9
66.5
69.2
4.39
4.06
84.4
80.5
68.9
-
-
-
Cosmos3-Nano [ 1 ]
16B
65.5
64.9
65.2
3.86
4.49
66.3
90.2
60.4
-
-
-
HunyuanVideo-1.5 [ 15 ]
8.3B
67.1
65.4
66.3
3.84
4.57
67.9
92.4
62.9
79.0
83.2
62.1
UniVideo [ 43 ]
13B
65.7
64.0
64.9
3.85
4.57
68.2
92.6
63.1
83.5
84.4
80.0
+ PhyT2V [ 45 ]
13B
63.5
67.1
65.3
3.96
4.52
69.7
91.2
64.3
83.1
84.1
79.1
Table 1: Evaluation on physics-focused benchmarks, i.e . , PhyGround [ 24 ] , VideoPhy-2 [ 3 ] ; and general quality benchmark, i.e . , VBench [ 11 ] . SA and PC denote Semantic Adherence and Physical Commonsense. We highlight the best and second best .
Figure 5: Qualitative results of PhysOPSD . More results are shown in Figures 14 , 15 , 16 and 17 .
Figure 6: Privilege context analysis. Executable coding drafts yield superior physical adherence over text and real videos.
Figure 7: Structured physical privilege analysis. Both ( Left ) the stepped dual-level hierarchy and ( Right ) parameter-controlled paired variations consistently yield superior physical adherence over their uncoupled counterparts.
Figure 8: Data scaling analysis. PhysOPSD delivers physical improvements even in low-data regimes and scales steadily.
Appendix figures & tables10 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 9: Comparison of PhysOPSD-Engine vs. Codex-Harness.
Figure 10
Figure 10: Distribution of PhysOPSD-3K .
Figure 11: Case study of PhysOPSD on rigid-body and contact dynamics. ( Left ) Pendulum wave : perturbing the damping rate ( 0.018 vs . 1.4 s−1 ) yields sustained wave oscillations vs . rapid kinetic decay. ( Right ) Domino cascade : varying gravitational acceleration ( 0.981 vs . 9.81 m/s2 ) alters the collapse speed from sluggish, incomplete toppling to rapid, sequential falling.
Figure 12: Case study of PhysOPSD on elastic and deformable solid mechanics. ( Left ) Transverse spring wave : modulating transverse coupling stiffness ( 0.5 vs . 8 N/m ) shifts dynamic response from localized dispersion to cohesive wave propagation. ( Right ) Sponge compression : increasing actuator displacement ( 8 vs . 24 mm ) drives the deformation from moderate compaction to severe, non-linear flattening.
Figure 13: Case study of PhysOPSD on fluid dynamics. ( Left ) Fluid drainage : altering the container outlet radius ( 1.8 vs . 5.5 mm ) controls the discharge flux from a thin, laminar stream to a high-volume outflow. ( Right ) Coffee stirring : varying spoon angular velocity ( 3 vs . 14 rad/s ) transitions the fluid motion from a smooth spiral vortex to vigorous turbulent mixing and foam dispersion.
Figure 14: Qualitative comparisons of PhysOPSD with baselines.
Figure 15: Qualitative comparisons of PhysOPSD with baselines.
Physical consistency in video generation studies how to make and evaluate videos whose geometry, motion, interactions, and state changes obey real-world laws.
Generative video models achieve high visual fidelity but often violate basic physical principles, limiting reliability in real-world settings. Prior attempts to inject physics rely on conditioning: frame-level signals are domain-specific and short-horizon, while global text prompts are coarse and noisy, missing fine-grained dynamics. We present PhysVid, a physics-aware local conditioning scheme that operates over temporally contiguous chunks of frames. Each chunk is annotated with physics-grounded descriptions of states, interactions, and constraints, which are fused with the global prompt via chunk-aware cross-attention during training. At inference, we introduce negative physics prompts (descriptions of locally relevant law violations) to steer generation away from implausible trajectories. On VideoPhy, PhysVid improves physical commonsense scores by ≈33% over baseline video generators, and by up to ≈8% on VideoPhy2. These results show that local, physics-aware guidance substantially increases physical plausibility in generative video and marks a step toward physics-grounded video models.
State-of-the-art text-to-video (T2V) generators frequently violate physical laws despite high visual quality. We show this stems from insufficient physical constraints in prompts rather than model limitations: manually adding physics details reliably produces physically plausible videos, but requires expertise and does not scale. We present PhyPrompt, a two-stage reinforcement learning framework that automatically refines prompts for physically realistic generation. First, we fine-tune a large language model on a physics-focused Chain-of-Thought dataset to integrate principles like object motion and force interactions while preserving user intent. Second, we apply Group Relative Policy Optimization with a dynamic reward curriculum that initially prioritizes semantic fidelity, then progressively shifts toward physical commonsense. This curriculum achieves synergistic optimization: PhyPrompt-7B reaches 40.8% joint success on VideoPhy2 (8.6pp gain), improving physical commonsense by 11pp (55.8% to 66.8%) while simultaneously increasing semantic adherence by 4.4pp (43.4% to 47.8%). Remarkably, our curriculum exceeds single-objective training on both metrics, demonstrating compositional prompt discovery beyond conventional multi-objective trade-offs. PhyPrompt outperforms GPT-4o (+3.8% joint) and DeepSeek-V3 (+2.2%, 100× larger) using only 7B parameters. The approach transfers zero-shot across diverse T2V architectures (Lavie, VideoCrafter2, CogVideoX-5B) with up to 16.8% improvement, establishing that domain-specialized reinforcement learning with compositional curricula surpasses general-purpose scaling for physics-aware generation.
Modern video generation models can synthesize visually compelling and temporally coherent clips, yet controlling their physical behavior remains difficult with standard text and image conditions. The core challenge is a conditioning bottleneck: material response, contact interaction, deformation, and motion trajectory are continuous and relational physical cues that are hard to specify exhaustively in language but can be demonstrated naturally by video. We propose VIPER, a Visual In-Context Physics Reasoning framework for reference-guided image-to-video generation. Given a target image, a brief target prompt, and a reference video, VIPER treats the reference as a dense visual demonstration of the desired physical process rather than an appearance template. It uses a Multimodal Large Language Model (MLLM) to extract reference-derived physical cues and guide a pretrained image-to-video generator through a hierarchical training strategy, enabling physical behavior transfer while preserving the visual prior of the base generator. To support this setting, we construct VIPER-19K, a curated dataset with material, trajectory, and physical-impact annotations, together with filtered reference-target pairs. Experiments on an unseen validation set show that VIPER achieves stronger reference-video physical similarity and higher human preference than representative video generation and video-as-prompt baselines, while maintaining competitive general video quality. Qualitative results further demonstrate that VIPER can transfer reference-derived physical behavior to new target scenes without requiring carefully engineered prompts.
Tianxiao Chen, Hanmo Chen, Huajin Chen +3
Zhejiang University · vivo BlueImage Lab, vivo Mobile Communication Co., Ltd.