LTX 2.5 Quantized needs about 16.4 GB of memory at Q4_K_M (a 14.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 7 quants, from 7.9 GB (Q2_K) to 22.7 GB (Q8_0).
LTX 2.5 Quantized – 21B model optimized for 16 GB GPUs
LTX 2.5 Quantized is a 21 billion‑parameter Ltxv model, released by joeygambino. It is a quantised version of Lightricks’ LTX‑2.5, designed to run on a 16 GB GPU. The model supports text generation and multimodal generation in a single pass.
Good at
- Text generation
- Multimodal generation
- Efficient GPU usage
Limits
- No vision support
- Requires separate text encoder
- License type unknown
Best for
- Local inference on 16 GB cards
- Rapid prototyping of multimodal apps
- Low‑cost deployment
Who it’s for. Developers and researchers who need a 21 B Ltx model that fits on a 16 GB GPU. The quantised weights allow inference without a custom node in ComfyUI. It is suitable for users who can manage the separate text encoder requirement.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 7.9 GB | 9.4 GB | Too big | Get file → | |
| Q3_K_M | 10.6 GB | 12.4 GB | Too big | Get file → | |
| Q4_K_S | 12.9 GB | 15.0 GB | Too big | Get file → | |
| Q4_K_M | 14.2 GB | 16.4 GB | Too big | Get file → | |
| Q5_K_M | 15.9 GB | 18.3 GB | Too big | Get file → | |
| Q6_K | 17.7 GB | 20.4 GB | Too big | Get file → | |
| Q8_0 | 22.7 GB | 26.0 GB | Too big | Get file → |
For a 16 GB card, the Q3_K_M quant (10.6 GB) or Q4_K_S (12.93 GB) are suitable. Q3_K_M uses about 12.4 GB of VRAM at 8 k context, while Q4_K_S uses about 15 GB. If more VRAM is available, Q4_K_M (14.17 GB) or Q5_K_M (15.9 GB) provide higher quality.
What GPU do you need for LTX 2.5 Quantized?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: the best quant that fits is Q2_K (7.9 GB file, about 9.4 GB in use).
- 16 GB VRAM: the best quant that fits is Q4_K_S (12.9 GB file, about 15.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (17.7 GB file, about 20.4 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (22.7 GB file, about 26.0 GB in use).
Which LTX 2.5 Quantized quant should you choose?
Q4_K_M is the usual starting point for LTX 2.5 Quantized: it balances file size and quality well. Here it is a 14.2 GB file.
The Q8_0 file is 22.7 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q2_K at 7.9 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about LTX 2.5 Quantized
How much VRAM does LTX 2.5 Quantized need?
About 16.4 GB at Q4_K_M with an 8K context window. That is the 14.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run LTX 2.5 Quantized on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 9.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of LTX 2.5 Quantized GGUF?
It depends on the quant: Q2_K is 7.9 GB, Q3_K_M is 10.6 GB, Q4_K_S is 12.9 GB, Q4_K_M is 14.2 GB, Q5_K_M is 15.9 GB, Q6_K is 17.7 GB and 1 more.
Where can I download LTX 2.5 Quantized GGUF?
The files are on Hugging Face in the joeygambino/LTX-2.5-Quantized repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does LTX 2.5 Quantized use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run LTX 2.5 Quantized on a 16 GB GPU?
Yes, the Q3_K_M and Q4_K_S quantisations fit on a 16 GB GPU, assuming the text encoder is not resident simultaneously.
Does the model support vision input?
No, the model does not have vision support in its current GGUF form.
Where can I find the quantised weights?
The weights are available in the joeygambino/LTX-2.5-Quantized Hugging Face repository as .gguf files.