Recognition: unknown
Wan: Open and Advanced Large-Scale Video Generative Models
read the original abstract
This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.
This paper has not been read by Pith yet.
Forward citations
Cited by 60 Pith papers
-
AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillation
AnyFlow enables any-step video diffusion by distilling flow-map transitions over arbitrary time intervals with on-policy backward simulation.
-
TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking
TrackCraft3R is the first method to repurpose a video diffusion transformer as a feed-forward dense 3D tracker via dual-latent representations and temporal RoPE alignment, achieving SOTA performance with lower compute.
-
Efficient Training on Multiple Consumer GPUs with RoundPipe
RoundPipe achieves near-zero-bubble pipeline parallelism for LLM training on consumer GPUs by dynamically dispatching computation stages round-robin, yielding 1.48-2.16x speedups and enabling 235B model fine-tuning on...
-
ViPS: Video-informed Pose Spaces for Auto-Rigged Meshes
ViPS distills a compact, controllable distribution of valid joint configurations for any auto-rigged mesh from video diffusion priors, matching 4D-trained methods in plausibility while generalizing zero-shot to unseen...
-
PhysInOne: Visual Physics Learning and Reasoning in One Suite
PhysInOne is a new dataset of 2 million videos across 153,810 dynamic 3D scenes covering 71 physical phenomena, shown to improve AI performance on physics-aware video generation, prediction, property estimation, and m...
-
OP-GRPO: Efficient Off-Policy GRPO for Flow-Matching Models
OP-GRPO is the first off-policy GRPO method for flow-matching models that reuses trajectories via replay buffer and importance sampling corrections, matching on-policy performance with 34.2% of the training steps.
-
ActivityForensics: A Comprehensive Benchmark for Localizing Manipulated Activity in Videos
ActivityForensics is the first large-scale benchmark for temporally localizing activity-level forgeries in videos, paired with a diffusion-based baseline called TADiff.
-
Flow-GRPO: Training Flow Matching Models via Online RL
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
-
EntityBench: Towards Entity-Consistent Long-Range Multi-Shot Video Generation
EntityBench is a new benchmark with detailed per-shot entity schedules from real media, and the EntityMem baseline using persistent per-entity memory achieves the highest character fidelity with Cohen's d of +2.33.
-
MiVE: Multiscale Vision-language features for reference-guided video Editing
MiVE repurposes VLMs as multiscale feature extractors integrated into a unified self-attention Diffusion Transformer, achieving top human preference in reference-guided video editing.
-
HASTE: Training-Free Video Diffusion Acceleration via Head-Wise Adaptive Sparse Attention
HASTE delivers up to 1.93x speedup on Wan2.1 video DiTs via head-wise adaptive sparse attention using temporal mask reuse and error-guided per-head calibration while preserving video quality.
-
Coding Agent Is Good As World Simulator
A multi-agent framework generates and refines executable physics simulation code from prompts to create world models that enforce physical constraints, claiming superior accuracy and fidelity over video-based alternatives.
-
TeDiO: Temporal Diagonal Optimization for Training-Free Coherent Video Diffusion
TeDiO regularizes temporal diagonals in diffusion transformer attention maps to produce smoother video motion while keeping per-frame quality intact.
-
R-DMesh: Video-Guided 3D Animation via Rectified Dynamic Mesh Flow
R-DMesh generates high-fidelity 4D meshes aligned to video by disentangling base mesh, motion, and a learned rectification jump offset inside a VAE, then using Triflow Attention and rectified-flow diffusion.
-
Asymmetric Flow Models
Asymmetric Flow Modeling restricts noise prediction to a low-rank subspace for high-dimensional flow generation, reaching 1.57 FID on ImageNet 256x256 and new state-of-the-art pixel text-to-image performance via finet...
-
CRePE: Curved Ray Expectation Positional Encoding for Unified-Camera-Controlled Video Generation
CRePE supplies depth-aware positional distributions along curved rays for stable unified-camera control in frozen video DiT models.
-
Inline Critic Steers Image Editing
Inline Critic uses a learnable token to critique and steer a frozen image-editing model's intermediate layers during generation, delivering state-of-the-art results on GEdit-Bench, RISEBench, and KRIS-Bench.
-
CausalCine: Real-Time Autoregressive Generation for Multi-Shot Video Narratives
CausalCine enables real-time causal autoregressive multi-shot video generation via multi-shot training, content-aware memory routing for coherence, and distillation to few-step inference.
-
OmniNFT: Modality-wise Omni Diffusion Reinforcement for Joint Audio-Video Generation
OmniNFT introduces modality-wise advantage routing, layer-wise gradient surgery, and region-wise loss reweighting in an online diffusion RL framework to improve audio-video quality, alignment, and synchronization.
-
From Imagined Futures to Executable Actions: Mixture of Latent Actions for Robot Manipulation
MoLA infers a mixture of latent actions from generated future videos via modality-aware inverse dynamics models to improve robot manipulation policies.
-
MoCam: Unified Novel View Synthesis via Structured Denoising Dynamics
MoCam uses structured denoising dynamics in diffusion models to temporally decouple geometric alignment from appearance refinement, enabling unified novel view synthesis that outperforms prior methods on imperfect poi...
-
MoCam: Unified Novel View Synthesis via Structured Denoising Dynamics
MoCam unifies static and dynamic novel view synthesis by temporally decoupling geometric alignment and appearance refinement within the diffusion denoising process.
-
$h$-control: Training-Free Camera Control via Block-Conditional Gibbs Refinement
h-control introduces block-conditional pseudo-Gibbs refinement for training-free camera control in flow-matching video generators, achieving superior FVD scores on RealEstate10K and DAVIS benchmarks.
-
CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating
CaC is a hierarchical spatiotemporal concentrating reward model for video anomalies that reports 25.7% accuracy gains on fine-grained benchmarks and 11.7% anomaly reduction in generated videos via a new dataset and GR...
-
Single-Shot HDR Recovery via a Video Diffusion Prior
Single-shot HDR is achieved by conditioning a video diffusion model on an LDR input to generate an exposure bracket and fusing the bracket with per-pixel weights from a lightweight UNet.
-
HorizonDrive: Self-Corrective Autoregressive World Model for Long-horizon Driving Simulation
HorizonDrive enables stable long-horizon autoregressive driving simulation via anti-drifting teacher training with scheduled rollout recovery and teacher rollout distillation.
-
PhyGround: Benchmarking Physical Reasoning in Generative World Models
PhyGround is a new benchmark with curated prompts, a 13-law taxonomy, large-scale human annotations, and an open physics-specialized VLM judge for evaluating physical reasoning in generative video models.
-
TIE: Time Interval Encoding for Video Generation over Events
TIE derives a sinc-based interval encoding from temporal integrability and duration invariance principles, raising temporal constraint satisfaction from 77% to 96% on the OmniEvents dataset while preserving visual quality.
-
Progressive Photorealistic Simplification
Progressive semantic image simplification uses VLMs and a verifier to iteratively remove and inpaint scene elements while preserving photorealism, distilled into an image-to-video model for direct sequence prediction.
-
Relative Score Policy Optimization for Diffusion Language Models
RSPO interprets reward advantages as targets for relative log-ratios in dLLMs, calibrating noisy estimates to stabilize RLVR training and achieve strong gains on planning tasks with competitive math reasoning performance.
-
CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Models
CollabVR improves video reasoning performance by coupling vision-language models and video generation models in a closed-loop step-level collaboration that detects and repairs generation failures.
-
From Articulated Kinematics to Routed Visual Control for Action-Conditioned Surgical Video Generation
A kinematic-to-visual lifting paradigm combined with hierarchically routed control generates action-conditioned surgical videos with better faithfulness, fidelity, and efficiency.
-
NoiseGate: Learning Per-Latent Timestep Schedules as Information Gating in World Action Models
NoiseGate learns per-latent timestep schedules as an information-gating policy in diffusion-based world action models, yielding consistent gains on RoboTwin manipulation tasks.
-
Improved techniques for fine-tuning flow models via adjoint matching: a deterministic control pipeline
A new adjoint matching framework formulates flow model alignment as optimal control, enabling direct regression training and terminal-trajectory truncation for efficiency gains on models like SiT-XL and FLUX.
-
Sparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance
Sparkle supplies a large-scale dataset and benchmark for instruction-driven video background replacement, enabling models that generate more natural and temporally consistent new scenes than earlier approaches.
-
FreeSpec: Training-Free Long Video Generation via Singular-Spectrum Reconstruction
FreeSpec uses SVD-based spectral reconstruction to fuse global low-rank and local high-rank features, reducing content drift and preserving temporal dynamics in long video generation.
-
Eulerian Motion Guidance: Robust Image Animation via Bidirectional Geometric Consistency
Eulerian adjacent-frame motion guidance plus bidirectional geometric consistency improves training speed, temporal coherence, and artifact reduction in diffusion-based image animation.
-
Eulerian Motion Guidance: Robust Image Animation via Bidirectional Geometric Consistency
Eulerian adjacent-frame motion fields with bidirectional cycle consistency checks enable faster parallel training and fewer artifacts in diffusion model image animation compared to initial-frame Lagrangian guidance.
-
EA-WM: Event-Aware Generative World Model with Structured Kinematic-to-Visual Action Fields
EA-WM generates more accurate robot world rollouts by projecting actions as structured visual fields in camera space and using event-aware bidirectional fusion to better capture interaction dynamics.
-
Stream-R1: Reliability-Perplexity Aware Reward Distillation for Streaming Video Generation
Stream-R1 improves distillation of autoregressive streaming video diffusion models by adaptively weighting supervision with a reward model at both rollout and per-pixel levels.
-
AniMatrix: An Anime Video Generation Model that Thinks in Art, Not Physics
AniMatrix generates anime videos using a structured taxonomy of artistic production variables, dual-channel conditioning, a style-motion curriculum, and deformation-aware optimization to prioritize art over physics.
-
AniMatrix: An Anime Video Generation Model that Thinks in Art, Not Physics
AniMatrix generates anime videos using a production knowledge taxonomy, dual-channel conditioning, style-motion curriculum, and deformation-aware preference optimization, outperforming baselines in animator evaluation...
-
AniMatrix: An Anime Video Generation Model that Thinks in Art, Not Physics
AniMatrix generates anime videos by structuring artistic production rules into a controllable taxonomy and training the model to prioritize those rules over physical realism, achieving top scores from professional ani...
-
WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video Models
WorldJen is a new benchmark for generative video models that uses VLM-judged multi-dimensional Likert questionnaires validated against human preferences to achieve perfect tier agreement.
-
TMD-Bench: A Multi-Level Evaluation Paradigm for Music-Dance Co-Generation
TMD-Bench is a multi-level benchmark that measures music-dance co-generation quality including beat-level rhythmic synchronization, supported by a new dataset and Music Captioner, and shows commercial models lag in rh...
-
VAnim: Rendering-Aware Sparse State Modeling for Structure-Preserving Vector Animation
VAnim creates open-domain text-to-SVG animations via sparse state updates on a persistent DOM tree, identification-first planning, and rendering-aware RL with a new 134k-example benchmark.
-
Action Agent: Agentic Video Generation Meets Flow-Constrained Diffusion
Action Agent pairs LLM-driven video generation with a flow-constrained diffusion transformer to produce velocity commands, raising video success to 86% and delivering 64.7% real-world navigation on a Unitree G1 humanoid.
-
AsymTalker: Identity-Consistent Long-Term Talking Head Generation via Asymmetric Distillation
AsymTalker maintains identity consistency in long-term diffusion talking-head videos by encoding temporal references from a static image and training a student model under inference-like conditions via asymmetric dist...
-
Thinking in Text and Images: Interleaved Vision--Language Reasoning Traces for Long-Horizon Robot Manipulation
A multimodal transformer generates and caches interleaved text-image traces to guide closed-loop actions, achieving 92.4% success on LIBERO-Long and 95.5% average on LIBERO.
-
FieryGS: In-the-Wild Fire Synthesis with Physics-Integrated Gaussian Splatting
FieryGS integrates LLM-based material reasoning, volumetric combustion simulation, and a unified renderer with 3D Gaussian Splatting to generate physically plausible and user-controllable fire in in-the-wild scenes.
-
TripVVT: A Large-Scale Triplet Dataset and a Coarse-Mask Baseline for In-the-Wild Video Virtual Try-On
A new large-scale triplet dataset and diffusion transformer model using coarse human masks deliver improved video virtual try-on quality and generalization in challenging real-world conditions.
-
Being-H0.7: A Latent World-Action Model from Egocentric Videos
Being-H0.7 adds future-aware latent reasoning to direct VLA policies via dual-branch alignment on latent queries, matching world-model benefits at VLA efficiency.
-
YOSE: You Only Select Essential Tokens for Efficient DiT-based Video Object Removal
YOSE accelerates DiT video object removal up to 2.5x by using BVI for adaptive token selection and DiffSim to simulate unmasked token effects, while preserving visual quality.
-
Cutscene Agent: An LLM Agent Framework for Automated 3D Cutscene Generation
Cutscene Agent uses a multi-agent LLM system and a new toolkit for game engine control to automate end-to-end 3D cutscene generation, evaluated on the introduced CutsceneBench.
-
DiscreteRTC: Discrete Diffusion Policies are Natural Asynchronous Executors
Discrete diffusion policies support native asynchronous execution via unmasking for real-time chunking, delivering higher success rates and 0.7x inference cost versus flow-matching RTC on dynamic robotics benchmarks a...
-
OmniShotCut: Holistic Relational Shot Boundary Detection with Shot-Query Transformer
OmniShotCut treats shot boundary detection as structured relational prediction via a shot-query Transformer, uses fully synthetic transitions for training data, and releases OmniShotCutBench for evaluation.
-
Characterizing Vision-Language-Action Models across XPUs: Constraints and Acceleration for On-Robot Deployment
VLA models exhibit a compute-bound VLM phase followed by a memory-bound action phase on edge hardware; DP-Cache and V-AEFusion reduce redundancy and enable pipeline parallelism for up to 6x speedup on NPUs with margin...
-
MuSS: A Large-Scale Dataset and Cinematic Narrative Benchmark for Multi-Shot Subject-to-Video Generation
MuSS is a movie-derived dataset and benchmark that enables AI models to generate multi-shot videos with coherent narratives and preserved subject identity across shots.
-
MuSS: A Large-Scale Dataset and Cinematic Narrative Benchmark for Multi-Shot Subject-to-Video Generation
MuSS is a new movie-sourced dataset and benchmark that enables AI models to generate multi-shot videos with improved narrative coherence and subject identity preservation.
-
Hallo-Live: Real-Time Streaming Joint Audio-Video Avatar Generation with Asynchronous Dual-Stream and Human-Centric Preference Distillation
Hallo-Live achieves 20.38 FPS real-time text-to-audio-video avatar generation with 0.94s latency using asynchronous dual-stream diffusion and HP-DMD preference distillation, matching teacher model quality at 16x highe...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.