Stay ahead in Video AI
A weekly email digest curating the best video generation and editing papers โ with benchmark scores, open-source model links, and one-sentence summaries. Every Tuesday.
Free forever ยท 1,247 researchers subscribed ยท No spam
12
Issues published
78
Papers covered
43%
Avg. open rate
What's inside every issue
Benchmark Tables
Side-by-side VBench, FVD, and CLIPSIM scores so you can compare models at a glance.
Model Links
Direct links to GitHub repos and HuggingFace checkpoints โ no hunting required.
One-Line Summaries
Each paper distilled to a single sentence so you know if it's worth reading.
Filter by Type
Papers tagged as generation or editing, plus model size and dataset info.
Tuesday Delivery
Arrives every Tuesday morning โ start your week knowing the latest in video AI.
Free Tier
The weekly digest is always free. Pay only if you want archive access.
Latest Issue
View all issues โIssue #12 โ Sora-Level Open Models & New VBench Records ยท 2026-03-03
OpenSora v2: Scaling Video Diffusion to 1080p with 3D-VAE
Introduces a 3D variational autoencoder that enables 1080p video generation at 24fps with significantly reduced memory cost.
| Metric | Score | ฮ |
|---|---|---|
| VBench Total | 84.3 | +2.1 |
| FVD (UCF-101) | 198 | -47 |
| CLIPSIM | 0.312 | +0.008 |
| Temporal Consistency | 97.1 | +1.4 |
VideoEdit-XL: Instruction-Based Video Editing with Flow Matching
A flow-matching approach for instruction-guided video editing that preserves temporal coherence across 128-frame clips.
| Metric | Score | ฮ |
|---|---|---|
| Edit Accuracy | 91.2% | +3.8% |
| Temporal Consistency | 95.6 | +2.3 |
| CLIPSIM | 0.298 | |
| FVD | 312 | -28 |
CogVideoX-5B: Scaling Expert Mixture for Long Video Generation
Mixture-of-experts architecture generating coherent 60-second videos with consistent characters and scene transitions.
| Metric | Score | ฮ |
|---|---|---|
| VBench Total | 83.7 | +1.9 |
| FVD (UCF-101) | 211 | -35 |
| Subject Consistency | 96.8 | +2.1 |
| CLIPSIM | 0.307 | +0.005 |
Don't miss the next issue
Join 1,247 researchers and creators getting the best video AI papers every Tuesday.