CogVideoX: what the repository publishes
CogVideoX is the one family here whose licence changes with the model size: the 2B model is Apache 2.0 and the 5B carries a separate licence of the project's own. Its smallest model is also the lowest hardware bar in this register, from 4GB. As of 2026-09-12.
| Field | What is stated | Detail |
|---|---|---|
| Licence | Split by model size | 2B is Apache 2.0; 5B carries a separate CogVideoX licence |
| Size | 2B and 5B, carried in the model names | The size decides which licence applies as well |
| Hardware | From 4GB in FP16 on the 2B | From 5GB on the 5B, from 10GB on 1.5-5B |
| Resolution | 1360x768 on 1.5-5B | 720x480 on the 5B |
| Length | Up to 161 frames on 1.5-5B | Described as 5 or 10 seconds |
Inclusion rule. The same five fields for every family here. A field the repository does not state is recorded as not stated rather than estimated from the parameter count. Order. Fixed field order, identical on every family page.
1One family, two licences, and the difference is the useful part
A reader who checks the licence on the smaller model and assumes it covers the larger one gets it wrong. That is not an unusual arrangement in this field and it is the only case in this register where it happens inside a single project.
The register therefore records the licence per model rather than per family. A family-level cell would have to pick one and would be wrong for whoever was reading about the other.

2Four gigabytes is a different category of accessible
Every other hardware figure in this register is measured in tens of gigabytes and implies a rented card. Four gigabytes in FP16 is a laptop-class graphics card, which puts this model in reach of someone experimenting rather than someone provisioning.
What that buys is the smaller model at the lower resolution. The register keeps the figures attached to the model they were stated for, because the gap between the smallest and largest entry in this family is itself a factor of several.
3Length is stated in frames and translated by the vendor
Up to 161 frames on the larger current model, which the repository describes as 5 or 10 seconds. Frames are the honest unit here because the seconds depend on frame rate, and the vendor supplies both so the register carries both.
Against the sixty seconds another family in this register states, this is short. Against what most hosted services sell as a single generation, it is comparable.
4The same field, across every family
| Family | Licence |
|---|---|
| Allegro | apache-2.0 |
| CogVideoX | Split by model size |
| Cosmos Predict 2.5 | NVIDIA Open Model License |
| FramePack | Apache 2.0 |
| HunyuanVideo | tencent-hunyuan-community |
| LongCat-Video | MIT License |
| LTX-Video | Apache 2.0 |
| MAGI-1 | Apache License 2.0 |
| Mochi 1 | Apache 2.0 |
| Open-Sora | Apache 2.0 |
| Pyramid Flow | stabilityai-ai-community |
| SkyReels-V2 | skywork-license |
| Stable Video Diffusion | stable-video-diffusion-community |
| Step-Video-T2V | MIT |
| Wan | Apache 2.0 |
Inclusion rule. Families whose vendor has published weights openly. A family whose repository does not state this particular field keeps its row, marked hollow. Order. Alphabetical by family name.
| Family | Size |
|---|---|
| Allegro | VAE 175M and DiT 2.8B |
| CogVideoX | 2B and 5B, carried in the model names |
| Cosmos Predict 2.5 | 2,059,174,912 parameters |
| FramePack | 13B models |
| HunyuanVideo | Over 13 billion parameters |
| LongCat-Video | 13.6B parameters |
| LTX-Video | A 13B model |
| MAGI-1 | 24B and 4.5B models |
| Mochi 1 | 10 billion parameters |
| Open-Sora | An 11B model |
| Pyramid Flow | No parameter count recorded |
| SkyReels-V2 | 1.3B, 5B and 14B variants |
| Stable Video Diffusion | 2B params |
| Step-Video-T2V | 30 billion parameters |
| Wan | A 5B text-to-video and image-to-video model |
Inclusion rule. Families whose vendor has published weights openly. A family whose repository does not state this particular field keeps its row, marked hollow. Order. Alphabetical by family name.
| Family | Hardware |
|---|---|
| Allegro | 9.3G BF16 with cpu_offload |
| CogVideoX | From 4GB in FP16 on the 2B |
| Cosmos Predict 2.5 | 32.54 GB of GPU VRAM |
| FramePack | 6GB minimum for 1800 frames at 30fps |
| HunyuanVideo | 60GB peak at 720px1280px, 129 frames |
| LongCat-Video | No memory figure recorded |
| LTX-Video | No minimum stated |
| MAGI-1 | H100 or H800 x 8 for the 24B |
| Mochi 1 | About 60GB VRAM on a single GPU |
| Open-Sora | 60.3GB peak at 768px on one GPU |
| Pyramid Flow | No memory figure recorded |
| SkyReels-V2 | About 14.7GB peak on the 1.3B at 540P |
| Stable Video Diffusion | About 180s on an A100 80GB card |
| Step-Video-T2V | 78.55 GB peak at 768px768px204f |
| Wan | A consumer 4090 on the 5B release |
Inclusion rule. Families whose vendor has published weights openly. A family whose repository does not state this particular field keeps its row, marked hollow. Order. Alphabetical by family name.
- LicenceSplit: CogVideoX-2B is Apache 2.0 while CogVideoX-5B carries a separate CogVideoX licencediffers by model size inside one family
- HardwareFrom 4GB in FP16 for the 2B model, from 5GB in BF16 for the 5B, and from 10GB for CogVideoX1.5-5Bthe lowest figures the repository states
- Length and resolutionCogVideoX1.5-5B states 1360x768 and up to 161 frames, described as 5 or 10 secondsper model
5Open weights, hosted
Instead of local hardware for “CogVideoX”, SceneMixer runs hosted models; its guides describe the result:
Video is 16:9 landscape by default, with 9:16 vertical available; you set it once per project. Segments are typically 4 to 15 seconds, and some model tiers go to 30.
SceneMixer, AI Short Drama Generator — “What length and aspect ratio does it output?” (read 2026-10-13)
Typically 60–120 seconds, always 9:16 vertical, ending on a hook. SceneMixer's episode outlines are generated with that pacing, and setting a project to 9:16 makes every segment vertical (the default is 16:9).
SceneMixer, How to Make ReelShort-Style Dramas with AI — “How long is a ReelShort-style episode?” (read 2026-10-13)
6Sources
Read from github.com/THUDM on 2026-09-12. Hardware figures across every family are collected on requirements; the licence column is set out on licence and what counts as a published weight is on the method page. Previous family: HunyuanVideo. Next family: Mochi 1.