Efficient Video Diffusion Models: Advancements and Challenges

2026-04-17

AI Summary

Video diffusion models face significant deployment challenges in high-fidelity video generation due to high inference costs, complex spatiotemporal computation, and substantial memory overhead. This work presents the first systematic survey of efficient video diffusion models, introducing a unified taxonomy that encompasses four mainstream optimization paradigms: step distillation, efficient attention mechanisms, model compression, and caching with trajectory optimization—each aimed at reducing the number of function evaluations and lowering per-step computational cost. The study clarifies the evolutionary trajectory of algorithms and their core optimization objectives, while highlighting promising future directions such as quality preservation, hardware-aware co-design, and long-duration video generation. By offering a structured reference, this survey aims to facilitate standardized evaluation and practical deployment of efficient video generation technologies.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

Video diffusion models have rapidly become the dominant paradigm for high-fidelity generative video synthesis, but their practical deployment remains constrained by severe inference costs. Compared with image generation, video synthesis compounds computation across spatial-temporal token growth and iterative denoising, making attention and memory traffic major bottlenecks in real-world settings. This survey provides a systematic and deployment-oriented review of efficient video diffusion models. We propose a unified categorization that organizes existing methods into four classes of main paradigms, including step distillation, efficient attention, model compression, and cache/trajectory optimization. Building on this categorization, we respectively analyze algorithmic trends of these four paradigms and examine how different design choices target two core objectives: reducing the number of function evaluations and minimizing per-step overhead. Finally, we discuss open challenges and future directions, including quality preservation under composite acceleration, hardware-software co-design, robust real-time long-horizon generation, and open infrastructure for standardized evaluation. To the best of our knowledge, our work is the first comprehensive survey on efficient video diffusion models, offering researchers and engineers a structured overview of the field and its emerging research directions.

Problem

• video diffusion models
• inference cost
• computational bottleneck
• memory traffic
• efficient generation

Innovation

• efficient video diffusion
• step distillation
• efficient attention
• model compression
• cache optimization