mlsanityData to 7 Oct 2026

SANA-Sprint: One-Step Diffusion with Continuous-Time Consistency Distillation

arXiv 2503.09641 ↗First posted 12 Mar 2025cs.GRcited by 85

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

This paper presents SANA-Sprint, an efficient diffusion model for ultra-fast text-to-image (T2I) generation. SANA-Sprint is built on a pre-trained foundation model and augmented with hybrid distillation, dramatically reducing inference steps from 20 to 1-4. We introduce three key innovations: (1) We propose a training-free approach that transforms a pre-trained flow-matching model for continuous-time consistency distillation (sCM), eliminating costly training from scratch and achieving high training efficiency. Our hybrid distillation strategy combines sCM with latent adversarial distillation (LADD): sCM ensures alignment with the teacher model, while LADD enhances single-step generation fidelity. (2) SANA-Sprint is a unified step-adaptive model that achieves high-quality generation in 1-4 steps, eliminating step-specific training and improving efficiency. (3) We integrate ControlNet with SANA-Sprint for real-time interactive image generation, enabling instant visual feedback for user interaction. SANA-Sprint establishes a new Pareto frontier in speed-quality tradeoffs, achieving state-of-the-art performance with 7.59 FID and 0.74 GenEval in only 1 step - outperforming FLUX-schnell (7.94 FID / 0.71 GenEval) while being 10x faster (0.1s vs 1.1s on H100). It also achieves 0.1s (T2I) and 0.25s (ControlNet) latency for 1024 x 1024 images on H100, and 0.31s (T2I) on an RTX 4090, showcasing its exceptional efficiency and potential for AI-powered consumer applications (AIPC). Code and pre-trained models will be open-sourced.

Key facts. Cited by 85 papers in the map (data through 7 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 66 of its 78 references that are in the map. First posted on arXiv 12 Mar 2025.

Builds on

66 of 78 references.

Closest ancestors · 9

Same direction, newest first

  1. 2024SnapGen: Taming High-Resolution Text-to-Image Models for Mobile Devices with Efficient Architectures and Training12 Dec
  2. 2024Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding14 May
  3. 2024Playground v2.5: Three Insights towards Enhancing Aesthetic Quality in Text-to-Image Generation27 Feb
  4. 2023MobileDiffusion: Instant Text-to-Image Generation on Mobile Devices28 Nov
  5. 2023GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment17 Oct
  6. 2023SnapFusion: Text-to-Image Diffusion Model on Mobile Devices within Two Seconds1 Jun
  7. 2023Wuerstchen: An Efficient Architecture for Large-Scale Text-to-Image Diffusion Models1 Jun
  8. 2023Adding Conditional Control to Text-to-Image Diffusion Models10 Feb
  9. 2022eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers2 Nov

Other references · 57

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2024The Llama 3 Herd of Modelscited by 14,192
  4. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  5. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  6. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  7. 2022Scalable Diffusion Models with Transformerscited by 5,931
  8. 2022Classifier-Free Diffusion Guidancecited by 5,717
  9. 2022Flow Matching for Generative Modelingcited by 5,426
  10. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,290
  11. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  12. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
38 more
  1. 2022DreamFusion: Text-to-3D using 2D Diffusioncited by 3,006
  2. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994
  3. 2023Consistency Modelscited by 1,778
  4. 2019Root Mean Square Layer Normalizationcited by 1,425
  5. 2023ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillationcited by 1,207
  6. 2023ImageReward: Learning and Evaluating Human Preferences for Text-to-Image Generationcited by 1,010
  7. 2023One-step Diffusion with Distribution Matching Distillationcited by 978
  8. 2022Building Normalizing Flows with Stochastic Interpolantscited by 964
  9. 2023PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesiscited by 854
  10. 2023Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inferencecited by 826
  11. 2023Adversarial Diffusion Distillationcited by 741
  12. 2022On Distillation of Guided Diffusion Modelscited by 721
  13. 2024Improved Distribution Matching Distillation for Fast Image Synthesiscited by 648
  14. 2015Holistically-Nested Edge Detectioncited by 590
  15. 2023Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusioncited by 448
  16. 2023Analyzing and Improving the Training Dynamics of Diffusion Modelscited by 396
  17. 2021Knowledge Distillation in Iterative Generative Models for Improved Sampling Speedcited by 344
  18. 2024SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformerscited by 331
  19. 2024PixArt-Sigma: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generationcited by 326
  20. 2024Fast High-Resolution Image Synthesis with Latent Adversarial Diffusion Distillationcited by 276
  21. 2024Simplifying, Stabilizing and Scaling Continuous-Time Consistency Modelscited by 276
  22. 2017Geometric GANcited by 232
  23. 2023Diff-Instruct: A Universal Approach for Transferring Knowledge From Pre-trained Diffusion Modelscited by 232
  24. 2024Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generationcited by 211
  25. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Modelscited by 209
  26. 2022Diffusion-GAN: Training GANs with Diffusioncited by 203
  27. 2025SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformercited by 165
  28. 2024SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Modelscited by 161
  29. 2024PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Modelscited by 139
  30. 2024Lumina-Next: Making Lumina-T2X Stronger and Faster with Next-DiTcited by 130
  31. 2024Playground v3: Improving Text-to-Image Alignment with Deep-Fusion Large Language Modelscited by 113
  32. 2024Distilling Diffusion Models into Conditional GANscited by 78
  33. 2024Multistep Distillation of Diffusion Models via Moment Matchingcited by 76
  34. 2023Diffusion Models Without Attentioncited by 75
  35. 2023Non-Denoising Forward-Time Diffusionscited by 75
  36. 2024DiM: Diffusion Mamba for Efficient High-Resolution Image Synthesiscited by 72
  37. 2024One-Step Diffusion Distillation through Score Implicit Matchingcited by 68
  38. 2024Multistep Consistency Modelscited by 67

50 most cited of 57 All 57 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 30; 2026: 55.

85 papers.

Sort:

By area:

  1. 2025Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistencycited by 81
  2. 2025TIIF-Bench: How Does Your T2I Model Follow Your Instructions?cited by 64
  3. 2026InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulationcited by 48
  4. 2025Distribution Matching Distillation Meets Reinforcement Learningcited by 43
  5. 2025Noise Hypernetworks: Amortizing Test-Time Compute in Diffusion Modelscited by 35
  6. 2025Transition Models: Rethinking the Generative Learning Objectivecited by 32
  7. 2025SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillationcited by 29
  8. 2026Transition Matching Distillation for Fast Video Generationcited by 27
  9. 2026Diamond Maps: Efficient Reward Alignment via Stochastic Flow Mapscited by 26
  10. 2025Playing with Transformer at 30+ FPS via Next-Frame Diffusioncited by 18
  11. 2025R3D2: Realistic 3D Asset Insertion via Diffusion for Autonomous Driving Simulationcited by 17
  12. 2025TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flowscited by 17
38 more
  1. 2025Psi-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Modelscited by 15
  2. 2025Fast Text-to-Audio Generation with Adversarial Post-Trainingcited by 14
  3. 2025Normalized Attention Guidance: Universal Negative Guidance for Diffusion Modelscited by 12
  4. 2025Adversarial Flow Modelscited by 10
  5. 2026SANA-Streaming: Real-time Streaming Video Editing with Hybrid Diffusion Transformercited by 10
  6. 2025SD3.5-Flash: Distribution-Guided Distillation of Generative Flowscited by 9
  7. 2025It's Never Too Late: Noise Optimization for Collapse Recovery in Trained Diffusion Modelscited by 9
  8. 2025HierarchicalPrune: Position-Aware Compression for Large-Scale Diffusion Modelscited by 8
  9. 2025Semantic Visual Anomaly Detection and Reasoning in AI-Generated Imagescited by 7
  10. 2025Stream-DiffVSR: Low-Latency Streamable Video Super-Resolution via Auto-Regressive Diffusioncited by 7
  11. 2026Skywork UniPic 3.0: Unified Multi-Image Composition via Sequence Modelingcited by 7
  12. 2025Score Distillation of Flow Matching Modelscited by 6
  13. 2025Improved Training Technique for Shortcut Modelscited by 6
  14. 2025FastFLUX: Pruning FLUX with Block-wise Replacement and Sandwich Trainingcited by 5
  15. 2025Soft-Di[M]O: Improving One-Step Discrete Image Generation with Soft Embeddingscited by 5
  16. 2025Block Cascading: Training Free Acceleration of Block-Causal Video Modelscited by 5
  17. 2025Few-Step Distillation for Text-to-Image Generation: A Practical Guidecited by 5
  18. 2026Stable Audio 3cited by 5
  19. 2025STORK: Faster Diffusion And Flow Matching Sampling By Resolving Both Stiffness And Structure-Dependencecited by 4
  20. 2026Faster and Better Alignment for Flow Matching Models via Step-aware Advantagescited by 4
  21. 2026Tiny Inference-Time Scaling with Latent Verifierscited by 4
  22. 2026Woosh: A Sound Effects Foundation Modelcited by 4
  23. 2026Continuous-Time Distribution Matching for Few-Step Diffusion Distillationcited by 4
  24. 2025Multimodal Benchmarking and Recommendation of Text-to-Image Generation Modelscited by 3
  25. 2025Shortcutting Pre-trained Flow Matching Diffusion Models is Almost Free Lunchcited by 3
  26. 2026PyramidalWan: On Making Pretrained Video Model Pyramidal for Efficient Inferencecited by 3
  27. 2026SnapGen++: Unleashing Diffusion Transformers for Efficient High-Fidelity Image Generation on Edge Devicescited by 3
  28. 2026Aligning Few-Step Generative Models by Amortizing Sample-based Variational Inferencecited by 3
  29. 2026OctoT2I: A Self-Evolving Agentic Text-to-Image Routercited by 3
  30. 2026Curvature-Adaptive Consistency Flow Matching: Autonomous Trajectory Optimization via Reinforcement Learningcited by 3
  31. 2026MobileWan: Closing the Quality Gap for Mobile Video Diffusioncited by 3
  32. 2025A Continuous-Time Consistency Model for 3D Point Cloud Generationcited by 2
  33. 2025FraQAT: Quantization Aware Training with Fractional bitscited by 2
  34. 2025From Structure to Detail: Hierarchical Distillation for Efficient Diffusion Modelcited by 2
  35. 2026InverFill: One-Step Inversion for Enhanced Few-Step Diffusion Inpaintingcited by 2
  36. 2026Self-Adversarial One Step Generation via Condition Shiftingcited by 2
  37. 2026Extending One-Step Image Generation from Class Labels to Text via Discriminative Text Representationcited by 2
  38. 2026Beyond Binary Preference: Aligning Diffusion Models to Fine-grained Criteria by Decoupling Attributescited by 1

50 most cited of 85 All 85 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation
  2. 2024Self-Corrected Flow Distillation for Consistent One-Step and Few-Step Text-to-Image Generation
  3. 2026Improving Controllable Generation: Faster Training and Better Performance via x0-Supervision
  4. 2023SpeedUpNet: A Plug-and-Play Adapter Network for Accelerating Text-to-Image Diffusion Models
  5. 2026Efficient Text-to-Image Generation: An Adaptive Step Schedule Controller for Diffusion Models
  6. 2023SwiftBrush: One-Step Text-to-Image Diffusion Model with Variational Score Distillation
  7. 2023CoDi: Conditional Diffusion Distillation for Higher-Fidelity and Faster Image Generation
  8. 2023HiPA: Enabling One-Step Text-to-Image Diffusion Models via High-Frequency-Promoting Adaptation