How much VRAM does LTX-2.5 need?
LTX-2.5 needs about 21.6 GB at 8-bit, or 13.1 GB as a 4-bit GGUF, at its default output size.
- Type
- Video model
- Parameters
- 21B
- Text encoder
- Gemma 4 12B
- Released
- Aug 2026
- License
- LTX-2.x Community License
Memory at each precision
| Precision | Weights | Peak at 544p, 5 s |
|---|---|---|
| 16-bit BF16 | 39.1 GB | 40.4 GB |
| 8-bit FP8 · Q8_0 | 20.3 GB | 21.6 GB |
| 6-bit GGUF Q6_K | 16.1 GB | 17.5 GB |
| 5-bit GGUF Q5_K_M | 13.9 GB | 15.3 GB |
| 4-bit GGUF Q4_K_M · NVFP4 | 11.7 GB | 13.1 GB |
The Gemma 4 12B text encoder (11.1 GB at FP8) runs first and then waits in system RAM.
What runs it
| Machine | Best fit | Per clip |
|---|---|---|
| RTX 3060 12 GB | Too big | |
| RTX 4060 Ti 16 GB | 5-bit | 2.8 min |
| RTX 3090 | 8-bit | 1.7 min |
| RTX 4090 | 8-bit | 54 s |
| RTX 5090 | 8-bit | 47 s |
| RTX PRO 6000 Blackwell | 16-bit | 31 s |
| M4 Max 128 GB | 16-bit | 6.5 min |
| M5 Max 128 GB | 16-bit | 2.1 min |
| Ryzen AI Max+ 395 128 GB | 16-bit | 3.1 min |
| DGX Spark 128 GB | 16-bit | 1.8 min |
| M3 Ultra 256 GB | 16-bit | 3.8 min |
| M5 Ultra 512 GB | 16-bit | 85 s |
| H100 SXM | 16-bit | 26 s |
| DGX Station GB300 | 16-bit | 22 s |
544p, 5 s, 8 steps. Times are rough, plus or minus 30%.
Where to get it
Questions
How much VRAM does LTX-2.5 need?
LTX-2.5 needs about 21.6 GB at 8-bit, or 13.1 GB as a 4-bit GGUF, at its default output size.
Can LTX-2.5 run on an RTX 4090?
Yes. At 8-bit it takes about 54 s per clip.
More models
Data checked Oct 2026.