Datasheet for local AI49 models98 GPUsData read 2026-09-25
No adsNo tracking
Check my GPU →

LTX-2.5 (22B) on 20 GB of VRAM

Video model · 22BAny 20 GB GPUData 2026-09-25
Runs
Yes, with a compressed file.

Download Q4_K_M (15.1 GB). With the model's working memory it needs about 19.9 GB, leaving 0.1 GB spare on 20 GB. Quality: good; some loss in fine detail and text. For better quality, Q5_K_M (16.8 GB) also runs, with about 1.6 GB spilling into system RAM — a little slower, still practical.

Best fileQ4_K_M
File size15.1 GB
VRAM needed~19.9 GB
System RAM48 GB+
Memory map · 20 GB GPU19.9 GB / 20 GB
010 GB20 GB
Weights Q4_K_M · 15.1 GBWorking memory · 4.0 GBReserve · 0.8 GBFree · 0.1 GB
calculated from real file sizes plus working memory. Not yet measured on this setup. How this works.

01What to download

PartFileFolderSize
ModelLTX-2.5-Distilled-Q4_K_M.gguf
Q4_K_M · realrebelai/LTX-2.5_GGUFs
models/unet15.1 GBDownload →
Text encodergemma4-12b-with-proj-ltx-2.5-comfy-int8-convrot.safetensors
Gemma 4 12B + LTX projection INT8 (convrot) · Lightricks/LTX-2.5
models/text_encoders15.4 GBDownload →
VAEltx-2.5-video-vae-bf16.safetensors
LTX-2.5 video VAE (DiffVAE) BF16 · Lightricks/LTX-2.5
models/vae1.5 GBDownload →
Also neededltx-2.5-audio-vae-bf16.safetensors
LTX-2.5 audio VAE + vocoder BF16 · Lightricks/LTX-2.5
models/vae0.4 GBDownload →
Total download · keep about the same free on disk32.3 GB

Alternatives: Gemma 4 12B + LTX projection BF16 (26.3 GB). The 16-bit text encoder is a little more faithful; the smaller one is chosen here because it loads faster and needs less RAM. Only for some workflows: LTX-2.5 latent spatial upscaler x2 (1.0 GB, official two-stage template); Gemma 4 E2B INT8 (prompt enhancer, optional) (5.2 GB, only if prompt enhancer enabled). Sizes read from Hugging Face (2026-09-25). “Download” links start the file directly; the file name links are the same files the official ComfyUI workflows use.

System RAM: 48 GB or more. ComfyUI keeps the model file, the text encoder and the VAE in system RAM and moves them to the GPU as needed. With this set of files that is about 32.3 GB, plus roughly 6 GB for Windows, ComfyUI and a browser: 38.3 GB in total. With less RAM it still runs, but Windows starts swapping to disk and loading gets very slow. calculated

02Every file, on 20 GB

FileSizeNeededOn this cardQualityDownload
16-bit
SAFETENSORS · Lightricks
42.0 GB46.8 GB26.8 GB too bigthe original weightsHugging Face →
Q8_0
GGUF · realrebelai
23.6 GB28.4 GB8.4 GB too bigpractically identical to the originalHugging Face →
Q6_K
GGUF · realrebelai
18.7 GB23.5 GB3.5 GB too bigvery close to the originalHugging Face →
Q5_K_M
GGUF · realrebelai
16.8 GB21.6 GBspills 1.6 GBclose; small differences in fine detailHugging Face →
Q4_K_M ←
GGUF · realrebelai
15.1 GB19.9 GBfits · 0.1 GB sparegood; some loss in fine detail and textHugging Face →
Q3_K_M
GGUF · realrebelai
11.5 GB16.3 GBfits · 3.7 GB sparenoticeable loss of detailHugging Face →
Q2_K
GGUF · realrebelai
8.8 GB13.6 GBfits · 6.4 GB spareheavy loss; a last resortHugging Face →

Sizes from the Hugging Face file listing, read 2026-09-25. “Needed” = file + 4 GB working memory for this model + 0.8 GB kept free for the system.

03The text encoder

Gemma 4 12B (custom LTX-2.5 build): 26.3 GB as 16-bit, 15.4 GB as INT8. ComfyUI encodes the prompt first and can push the encoder out of VRAM before sampling, so it does not have to fit together with the model. On 20 GB the INT8 encoder fits on its own, so prompt encoding stays fast.