Check my GPU →

LTX-2.3 (22B) on 12 GB of VRAM

Not entirely. Even Q2_K is 1.1 GB over 12 GB, so ComfyUI keeps part of it in system RAM. The March 2026 update of LTX-2, grown to 22B parameters.

Video model · 22BAny 12 GB GPUData 2026-10-01
Offload
Not entirely in VRAM, but it can still run.

Nothing fits entirely, but Q2_K (8.3 GB) overflows by only about 1.1 GB. ComfyUI keeps that part in system RAM automatically: slower than a full fit, but usable.

Best fileQ2_K
File size8.3 GB
VRAM needed~13.1 GB
System RAM32 GB+
Memory map · 12 GB GPU13.1 GB needed · 1.1 GB over 12 GB
07 GB13 GB
Weights Q2_K · 8.3 GBWorking memory · 4.0 GBReserve · 0.8 GBSpills to system RAM · 1.1 GB
calculated How this is counted.

01Other amounts of VRAM

6 GB Not practical8 GB Offload only10 GB Offload only16 GB Tight20 GB Runs24 GB Runs32 GB Runs well

02Every file, on 12 GB

FileSizeNeededOn this cardQualityDownload
16-bit
GGUF · unsloth
42.0 GB46.8 GB34.8 GB too bigthe original weightsHugging Face →
Q8_0
GGUF · unsloth
22.8 GB27.6 GB15.6 GB too bigpractically identical to the originalHugging Face →
Q6_K
GGUF · unsloth
17.8 GB22.6 GB10.6 GB too bigvery close to the originalHugging Face →
Q5_K_M
GGUF · unsloth
16.1 GB20.9 GB8.9 GB too bigclose; small differences in fine detailHugging Face →
Q4_K_M
GGUF · unsloth
14.3 GB19.1 GB7.1 GB too biggood; some loss in fine detail and textHugging Face →
Q3_K_M
GGUF · unsloth
10.8 GB15.6 GB3.6 GB too bignoticeable loss of detailHugging Face →
Q2_K ←
GGUF · unsloth
8.3 GB13.1 GBspills 1.1 GBheavy loss; a last resortHugging Face →

“Needed” adds 4 GB of working memory for LTX-2.3 and a 0.8 GB reserve. Sizes 2026-10-01.

03What to download

PartFileFolderSize
Modelltx-2.3-22b-dev-Q2_K.gguf
Q2_K · unsloth/LTX-2.3-GGUF
models/unet8.3 GBDownload →
Text encodergemma_3_12B_it_fp8_scaled.safetensors
Gemma 3 12B FP8 (scaled) · Comfy-Org/ltx-2
models/text_encoders13.2 GBDownload →
VAEltx-2.3-22b-dev_video_vae.safetensors
LTX-2.3 video VAE · unsloth/LTX-2.3-GGUF
models/vae1.5 GBDownload →
Also neededltx-2.3-22b-dev_audio_vae.safetensors
LTX-2.3 audio VAE · unsloth/LTX-2.3-GGUF
models/vae0.4 GBDownload →
Also neededltx-2.3-22b-dev_embeddings_connectors.safetensors
LTX-2.3 text embeddings connectors (projection) · unsloth/LTX-2.3-GGUF
models/text_encoders2.3 GBDownload →
Total download · keep about the same free on disk25.6 GB

Alternatives: Gemma 3 12B FP4 mixed (9.4 GB); Gemma 3 12B BF16 (24.4 GB). FP4 files are listed only as alternatives unless the GPU is an RTX 50 (Blackwell). Only for some workflows: LTX-2.3 spatial upscaler x2 1.1 (1.0 GB, official two-stage template). Sizes from Hugging Face, 2026-10-01.

04The text encoder

Take the smallest Gemma 3 12B file (9.4 GB): it is the only one that fits in 12 GB on its own. Sizes: 16-bit 24.4 GB, FP8 13.2 GB, FP4 mixed 9.4 GB. Why the encoder does not count against VRAM →