Research map / Video / Video generation
The Best of Both Worlds: Integrating Language Models and Diffusion Models for Video Generation
Aoxiong Yin, Xu Tan, Kai Shen, Yichong Leng, Xinyu Zhou, Juncheng Li, Siliang Tang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add the semantic video tokenizer (vs ARLON-like VAE-feature LLM) (ablation) | LanDiff · VBench | Total Score: 82.31 -> 85.43 | Improved |
| add the semantic video tokenizer (ablation) | LanDiff · VBench | Semantic Score: 77.27 -> 82.61 | Improved |
| add the semantic video tokenizer (ablation) | LanDiff · VBench | Quality Score: 83.58 -> 86.13 | Improved |
| add classifier-free guidance for text (ablation) | LanDiff · VBench | Total Score: 81.06 -> 85.43 | Improved |
| add classifier-free guidance for text (ablation) | LanDiff · VBench | Semantic Score: 73.13 -> 82.61 | Improved |
| add classifier-free guidance for text (ablation) | LanDiff · VBench | Quality Score: 83.04 -> 86.13 | Improved |
| combine LLM semantic tokens with diffusion instead of diffusion alone (same data) | LanDiff vs 7B DiT · VBench | Total Score: 81.85 -> 85.43 | Improved |
| combine LLM semantic tokens with diffusion instead of diffusion alone (same data) | LanDiff vs 7B DiT · VBench | Dynamic Degree: 51.02 -> 92.71 | Improved |
| use hybrid LLM+diffusion at 5B vs CogVideoX-5B | LanDiff · VBench | Total Score: 81.61 -> 85.43 | Improved |
| use hybrid 5B model vs larger HunyuanVideo 13B | LanDiff · VBench | Total Score: 83.24 -> 85.43 | Improved |
| use hybrid model vs ARLON for long video | LanDiff · VBench-Long | Total Score: 65.09 -> 68.34 | Improved |
| use LanDiff vs ARLON for long video | LanDiff · VBench-Long | Dynamic Degree: 50.42 -> 72.86 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.
- CogVideoX-5B — CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer (ICLR 2025) · better on 4 of 4 shared results
- StreamingT2V — StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025) · better on 1 of 2 shared results
- ARLON — ARLON: Boosting Diffusion Transformers with Autoregressive Models for Long Video Generation (ICLR 2025) · better on 2 of 2 shared results
- FreeNoise — FreeNoise: Tuning-Free Longer Video Diffusion via Noise Rescheduling (ICLR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SEINE: Short-to-Long Video Diffusion Model for Generative Transition and Prediction (ICLR 2024 · similarity 0.90)
- A Recipe for Scaling up Text-to-Video Generation with Text-free Videos (CVPR 2024 · similarity 0.89)
- Language Model Beats Diffusion - Tokenizer is key to visual generation (ICLR 2024 · similarity 0.89)
- Mimir: Improving Video Diffusion Models for Precise Text Understanding (CVPR 2025 · similarity 0.89)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.89)
- OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation (ICLR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
