# TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times arXiv: https://arxiv.org/abs/2512.16093 · HTML: https://mlsanity.com/paper/2512.16093/ ## Key facts - Cited by 33 papers in the map (data through 14 Oct 2026). - Placed in Video generation, one of 14 directions in Diffusion models. - Builds on 6 of its 7 references that are in the map. - First posted on arXiv 18 Dec 2025. ## Placed in - Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/) ## Builds on (most cited first, 6 of 6) - [SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration](https://mlsanity.com/paper/2410.02367/) (2024) - [SageAttention2: Efficient Attention with Thorough Outlier Smoothing and Per-thread INT4 Quantization](https://mlsanity.com/paper/2411.10958/) (2024) - [Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency](https://mlsanity.com/paper/2510.08431/) (2025) - [SageAttention3: Microscaling FP4 Attention for Inference and An Exploration of 8-Bit Training](https://mlsanity.com/paper/2505.11594/) (2025) - [SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention](https://mlsanity.com/paper/2509.24006/) (2025) - [SageAttention2++: A More Efficient Implementation of SageAttention2](https://mlsanity.com/paper/2505.21136/) (2025) ## Built on by (most cited first, 30 of 33) - [Helios: Real Real-Time Long Video Generation Model](https://mlsanity.com/paper/2603.04379/) (2026) - [Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models](https://mlsanity.com/paper/2606.25473/) (2026) - [SLA2: Sparse-Linear Attention with Learnable Routing and QAT](https://mlsanity.com/paper/2602.12675/) (2026) - [SpargeAttention2: Trainable Sparse Attention via Hybrid Top-k+Top-p Masking and Distillation Fine-Tuning](https://mlsanity.com/paper/2602.13515/) (2026) - [KernelBenchX: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels](https://mlsanity.com/paper/2605.04956/) (2026) - [SageBwd: A Trainable Low-bit Attention](https://mlsanity.com/paper/2603.02170/) (2026) - [Efficient Video Diffusion Models: Advancements and Challenges](https://mlsanity.com/paper/2604.15911/) (2026) - [BOute: Cost-Efficient LLM Serving with Heterogeneous LLMs and GPUs via Multi-Objective Bayesian Optimization](https://mlsanity.com/paper/2602.10729/) (2026) - [MonarchRT: Efficient Attention for Real-Time Video Generation](https://mlsanity.com/paper/2602.12271/) (2026) - [Proprio: Latent Self-Scoring and Inference-Time Refinement for Physically Plausible Video Generation](https://mlsanity.com/paper/2605.28230/) (2026) - [Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms](https://mlsanity.com/paper/2603.28489/) (2026) - [Taming I2V models for Image HOI Editing: A Cognitive Benchmark and Agentic Self-Correcting Framework](https://mlsanity.com/paper/2606.19073/) (2026) - [Vidu S2: Real-Time Interactive, Editable, and Spatial Video Generation](https://mlsanity.com/paper/2609.11638/) (2026) - [Vidu S1: A Real-Time Interactive Video Generation Model](https://mlsanity.com/paper/2607.03118/) (2026) - [Diffusing to Coordinate: Efficient Online Multi-Agent Diffusion Policies](https://mlsanity.com/paper/2602.18291/) (2026) - [CachedSearch: Training-Free Cached Exploration for Test-Time Search in Video Diffusion](https://mlsanity.com/paper/2607.23159/) (2026) - [Speculative Decoding for Autoregressive Video Generation](https://mlsanity.com/paper/2604.17397/) (2026) - [OSP-Next: Efficient High-Quality Video Generation with Sparse Sequence Parallelism, HiF8 Quantization, and Reinforcement Learning](https://mlsanity.com/paper/2605.28691/) (2026) - [KVAE: Family of Tokenizers for Multimodal Generative Models](https://mlsanity.com/paper/2608.05798/) (2026) - [FrescoDiffusion: 4K Image-to-Video with Prior-Regularized Tiled Diffusion](https://mlsanity.com/paper/2603.17555/) (2026) - [6Bit-Diffusion: Inference-Time Mixed-Precision Quantization for Video Diffusion Models](https://mlsanity.com/paper/2603.18742/) (2026) - [DiV-INR: Extreme Low-Bitrate Diffusion Video Compression with INR Conditioning](https://mlsanity.com/paper/2604.08329/) (2026) - [Reward Lightning: Fast Video Generation via Homologous Preference Distillation](https://mlsanity.com/paper/2607.03960/) (2026) - [Optimal Transport Q-Learning for Flow Policy Steering and Acceleration](https://mlsanity.com/paper/2607.06262/) (2026) - [Dynamic-in-Few-Step: Unifying Dynamic Computation and Few-Step Distillation for Efficient Video Generation](https://mlsanity.com/paper/2607.06631/) (2026) - [Kaleido: Algorithm-Hardware Co-Design for Video Diffusion Transformers by Exploiting Latent Space Correlations](https://mlsanity.com/paper/2607.13770/) (2026) - [OSVE: One Step Video Editing with One Step Diffusion Models](https://mlsanity.com/paper/2607.19895/) (2026) - [TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation](https://mlsanity.com/paper/2608.24674/) (2026) - [DSAQuant: Denoising-Stage-Aligned Quantization-Aware Training for Video Generation](https://mlsanity.com/paper/2609.04031/) (2026) - [SparkDiffusion: Mitigating the High-Sparsity Trap --- A Unified Framework for up to 265× Single-GPU Acceleration of Visual Generation](https://mlsanity.com/paper/2609.23153/) (2026) ## Abstract (arXiv) We introduce TurboDiffusion, a video generation acceleration framework that can speed up end-to-end diffusion generation by 100-200x while maintaining video quality. TurboDiffusion mainly relies on several components for acceleration: (1) Attention acceleration: TurboDiffusion uses low-bit SageAttention and trainable Sparse-Linear Attention (SLA) to speed up attention computation. (2) Step distillation: TurboDiffusion adopts rCM for efficient step distillation. (3) W8A8 quantization: TurboDiffusion quantizes model parameters and activations to 8 bits to accelerate linear layers and compress the model. In addition, TurboDiffusion incorporates several other engineering optimizations. We conduct experiments on the Wan2.2-I2V-14B-720P, Wan2.1-T2V-1.3B-480P, Wan2.1-T2V-14B-720P, and Wan2.1-T2V-14B-480P models. Experimental results show that TurboDiffusion achieves 100-200x speedup for video generation even on a single RTX 5090 GPU, while maintaining comparable video quality. The GitHub repository, which includes model checkpoints and easy-to-use code, is available at https://github.com/thu-ml/TurboDiffusion.