COpen
CogVideoX-5B
Tsinghua / Zhipu AI
Lightweight entry point. 12GB GPU, Apache 2.0. Most accessible for experimentation.
SOpen
LTX-2.3
Lightricks
The open-source flagship. First open model to close the gap with proprietary leaders.
Pick CogVideoX-5B if…
You want budget prototyping, research, or motion-heavy short clips.
Pick LTX-2.3 if…
You want open-source audio-video, or studios needing IP control.
Specifications
Maker
Tsinghua / Zhipu AI
Lightricks
Source Type
Open Source
Open Source
License
Apache 2.0
Apache 2.0 (<$10M rev)
Architecture
Expert Transformer
DiT (22B) + Rebuilt VAE
Parameters
5B
22B
Max Resolution
720p (1360x768)
4K
Max Duration
6s
20s
FPS
8
Up to 50
Native Audio
No
Yes
ComfyUI Support
Yes
Yes
Fine-tunable
Yes
Yes
Min VRAM
12GB
32GB (full) / 12GB (distilled)
Cost / Second
Self-host
$0.04
Inputs
T2V
T2V, I2V, V2V, Audio-cond
On Floyo
No
Yes
Strengths & Trade-offs
CogVideoX-5B
Strengths
- +Lightweight 12GB
- +Apache 2.0
- +LoRA + DDIM Inverse
- +ModelScope integration
Trade-offs
- -Low FPS (8fps)
- -6s max
- -older architecture
Best For
- →Budget prototyping
- →research
- →motion-heavy short clips
LTX-2.3
Strengths
- +22B params
- +true 4K at 50fps
- +first open model with synced audio
- +rebuilt VAE
- +native portrait
Trade-offs
- -Full 4K needs 48GB
- -dialogue lip-sync inconsistent
- -in-scene text flaky
Best For
- →Local 4K production
- →open-source audio-video
- →studios needing IP control
Run these models on Floyo
Browser-based ComfyUI. No setup, no GPU required.
I2VLTX-2.3
LTX 2.3 Pro Image to Video
1.0k runs
A2VLTX-2.3
LTX 2.3 Audio to Video
124 runs
T2VLTX-2.3
LTX 2.3 Pro Text to Video
101 runs
T2VLTX-2.3
LTX 2.3 T2V (Community)
47 runs