mlsanityData to 7 Oct 2026

LongLive: Real-time Interactive Long Video Generation

arXiv 2509.22622 ↗First posted 26 Sep 2025cs.CVcited by 237

Not on the map. Why?

Cited by 94 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We present LongLive, a frame-level autoregressive (AR) framework for real-time and interactive long video generation. Long video generation presents challenges in both efficiency and quality. Diffusion and Diffusion-Forcing models can produce high-quality videos but suffer from low efficiency due to bidirectional attention. Causal attention AR models support KV caching for faster inference, but often degrade in quality on long videos due to memory challenges during long-video training. In addition, beyond static prompt-based generation, interactive capabilities, such as streaming prompt inputs, are critical for dynamic content creation, enabling users to guide narratives in real time. This interactive requirement significantly increases complexity, especially in ensuring visual consistency and semantic coherence during prompt transitions. To address these challenges, LongLive adopts a causal, frame-level AR design that integrates a KV-recache mechanism that refreshes cached states with new prompts for smooth, adherent switches; streaming long tuning to enable long video training and to align training and inference (train-long-test-long); and short window attention paired with a frame-level attention sink, shorten as frame sink, preserving long-range consistency while enabling faster generation. With these key designs, LongLive fine-tunes a 1.3B-parameter short-clip model to minute-long generation in just 32 GPU-days. At inference, LongLive sustains 20.7 FPS on a single NVIDIA H100, achieves strong performance on VBench in both short and long videos. LongLive supports up to 240-second videos on a single H100 GPU. LongLive further supports INT8-quantized inference with only marginal quality loss.

Key facts. Cited by 237 papers in the map (data through 7 Oct 2026). Builds on 44 of its 49 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

44 of 49 references.

References · 44

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2022Scalable Diffusion Models with Transformerscited by 5,931
  3. 2025Wan: Open and Advanced Large-Scale Video Generative Modelscited by 2,536
  4. 2024CogVideoX: Text-to-Video Diffusion Models with An Expert Transformercited by 2,198
  5. 2024Qwen2 Technical Reportcited by 2,076
  6. 2023VBench: Comprehensive Benchmark Suite for Video Generative Modelscited by 1,755
  7. 2024HunyuanVideo: A Systematic Framework For Large Video Generative Modelscited by 1,597
  8. 2023One-step Diffusion with Distribution Matching Distillationcited by 978
  9. 2024Improved Distribution Matching Distillation for Fast Image Synthesiscited by 648
  10. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 634
  11. 2025Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusioncited by 569
  12. 2024LTX-Video: Realtime Video Latent Diffusioncited by 549
32 more
  1. 2022Phenaki: Variable Length Video Generation From Open Domain Textual Descriptioncited by 475
  2. 2024From Slow Bidirectional to Fast Autoregressive Video Diffusion Modelscited by 401
  3. 2022Latent Video Diffusion Models for High-Fidelity Long Video Generationcited by 385
  4. 2024Pyramidal Flow Matching for Efficient Video Generative Modelingcited by 354
  5. 2023LAVIE: High-Quality Video Generation with Cascaded Latent Diffusion Modelscited by 349
  6. 2025MAGI-1: Autoregressive Video Generation at Scalecited by 262
  7. 2023LongLoRA: Efficient Fine-tuning of Long-Context Large Language Modelscited by 243
  8. 2024StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Textcited by 242
  9. 2023SEINE: Short-to-Long Video Diffusion Model for Generative Transition and Predictioncited by 231
  10. 2025SkyReels-V2: Infinite-length Film Generative Modelcited by 218
  11. 2024Autoregressive Video Generation without Vector Quantizationcited by 180
  12. 2023FreeNoise: Tuning-Free Longer Video Diffusion via Noise Reschedulingcited by 173
  13. 2023NUWA-XL: Diffusion over Diffusion for eXtremely Long Video Generationcited by 170
  14. 2025History-Guided Video Diffusioncited by 165
  15. 2024SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Modelscited by 161
  16. 2025Frame Context Packing and Drift Prevention in Next-Frame-Prediction Video Diffusion Modelscited by 126
  17. 2024VidProM: A Million-scale Real Prompt-Gallery Dataset for Text-to-Video Diffusion Modelscited by 124
  18. 2025Long-Context Autoregressive Video Modeling with Next-Frame Predictioncited by 119
  19. 2025Long Context Tuning for Video Generationcited by 110
  20. 2025One-Minute Video Generation with Test-Time Trainingcited by 100
  21. 2024FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attentioncited by 87
  22. 2024The Matrix: Infinite-Horizon World Generation with Real-Time Moving Controlcited by 79
  23. 2025Yume: An Interactive World Generation Modelcited by 73
  24. 2024SEED-Story: Multimodal Long Story Generation with Large Language Modelcited by 70
  25. 2025RIFLEx: A Free Lunch for Length Extrapolation in Video Diffusion Transformerscited by 63
  26. 2025StreamDiT: Real-Time Streaming Text-to-Video Generationcited by 43
  27. 2025MoCha: Towards Movie-Grade Talking Character Synthesiscited by 31
  28. 2025Taming Teacher Forcing for Masked Autoregressive Video Generationcited by 29
  29. 2023Denoising Diffusion Step-aware Modelscited by 24
  30. 2025Lumos-1: On Autoregressive Video Generation with Discrete Diffusion from a Unified Model Perspectivecited by 22
  31. 2025LongVie: Multimodal-Guided Controllable Ultra-Long Video Generationcited by 21
  32. 2025FreeLong++: Training-Free Long Video Generation via Multi-band SpectralFusioncited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 29; 2026: 208.

237 papers.

Sort:

By area:

  1. 2025Self-Forcing++: Towards Minute-Scale High-Quality Video Generationcited by 176
  2. 2026Causal Forcing: Autoregressive Diffusion Distillation Done Right for High-Quality Real-Time Interactive Video Generationcited by 159
  3. 2025WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modelingcited by 152
  4. 2026Advancing Open-source World Modelscited by 144
  5. 2025SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformercited by 91
  6. 2025RELIC: Interactive Video World Model with Long-Horizon Memorycited by 89
  7. 2025Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rolloutcited by 78
  8. 2025Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillationcited by 75
  9. 2025Deep Forcing: Training-Free Long Video Generation with Deep Sink and Participative Compressioncited by 72
  10. 2025Stable Video Infinity: Infinite-Length Video Generation with Error Recyclingcited by 68
  11. 2026Helios: Real Real-Time Long Video Generation Modelcited by 56
  12. 2026SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformercited by 50
38 more
  1. 2026Context Forcing: Consistent Autoregressive Video Generation with Long Contextcited by 47
  2. 2025Memory Forcing: Spatio-Temporal Memory for Consistent Scene Generation on Minecraftcited by 46
  3. 2026WoVR: World Models as Reliable Simulators for Post-Training VLA Policies with RLcited by 46
  4. 2025Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidancecited by 45
  5. 2025End-to-End Training for Autoregressive Video Diffusion via Self-Resamplingcited by 45
  6. 2026Causal Forcing++: Scalable Few-Step Autoregressive Diffusion Distillation for Real-Time Interactive Video Generationcited by 44
  7. 2025VideoSSM: Autoregressive Long Video Generation with Hybrid State-Space Memorycited by 39
  8. 2026MOVA: Towards Scalable and Synchronized Video-Audio Generationcited by 39
  9. 2026Infinite Worlds with Versatile Interactionscited by 37
  10. 2026DreamX-World 1.0: A General-Purpose Interactive World Modelcited by 33
  11. 2026LoL: Longer than Longer, Scaling Video Generation to Hourcited by 29
  12. 2026Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusioncited by 28
  13. 2025StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generationcited by 27
  14. 2026LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generationcited by 25
  15. 2025StreamAvatar: Streaming Diffusion Models for Real-Time Interactive Human Avatarscited by 24
  16. 2026AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillationcited by 23
  17. 2026Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Modelscited by 23
  18. 2026VideoMemory: Toward Consistent Video Generation via Memory Integrationcited by 22
  19. 2026ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytellingcited by 22
  20. 2025FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenizationcited by 21
  21. 2026MemRoPE: Training-Free Infinite Video Generation via Evolving Memory Tokenscited by 20
  22. 2026Relax Forcing: Relaxed KV-Memory for Consistent Long Video Generationcited by 20
  23. 2026RealWonder: Real-Time Physical Action-Conditioned Video Generationcited by 18
  24. 2026Self Gradient Forcing: Native Long Video Extrapolationcited by 16
  25. 2025DC-VideoGen: Efficient Video Generation with Deep Compression Video Autoencodercited by 15
  26. 2026Efficient Autoregressive Video Diffusion with Dummy Headcited by 15
  27. 2026MBench: A Comprehensive Benchmark on Memory Capability for Video World Modelscited by 15
  28. 2026LIVE: Long-horizon Interactive Video World Modelingcited by 14
  29. 2026Anchor Forcing: Anchor Memory and Tri-Region RoPE for Interactive Streaming Video Diffusioncited by 14
  30. 2026PackForcing: Short Video Training Suffices for Long Video Sampling and Long Context Inferencecited by 14
  31. 2026WorldKV: Efficient World Memory with World Retrieval and Compressioncited by 14
  32. 2025BAgger: Backwards Aggregation for Mitigating Drift in Autoregressive Video Diffusion Modelscited by 13
  33. 2026Past- and Future-Informed KV Cache Policy with Salience Estimation in Autoregressive Video Diffusioncited by 13
  34. 2026AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memoriescited by 13
  35. 2026HiAR: Efficient Autoregressive Long Video Generation via Hierarchical Denoisingcited by 13
  36. 2026WildWorld: A Large-Scale Dataset for Dynamic World Modeling with Actions and Explicit State toward Generative ARPGcited by 13
  37. 2026One-Forcing: Towards Stable One-Step Autoregressive Video Generationcited by 13
  38. 2026ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPUcited by 13

50 most cited of 237 All 237 (JSON)