LTX-2 (19B) VRAM requirements
Lightricks' open audio+video model. Generates sound together with the picture. Its text encoder is Gemma 3 12B.
TypeVideo
Parameters19B
Fits entirely from13 GB
8-bit or better from26 GB
01The files, and how much VRAM each needs
| File | Size | Needed | Min. VRAM | Quality | Source |
|---|---|---|---|---|---|
| 16-bit | 37.8 GB | 42.6 GB | 43 GB | the original weights | unsloth/LTX-2-GGUF → |
| Q8_0 | 20.4 GB | 25.2 GB | 26 GB | practically identical to the original | unsloth/LTX-2-GGUF → |
| Q6_K | 16.0 GB | 20.8 GB | 21 GB | very close to the original | unsloth/LTX-2-GGUF → |
| Q5_K_M | 14.3 GB | 19.1 GB | 20 GB | close; small differences in fine detail | unsloth/LTX-2-GGUF → |
| Q4_K_M | 12.8 GB | 17.6 GB | 18 GB | good; some loss in fine detail and text | unsloth/LTX-2-GGUF → |
| Q3_K_M | 10.1 GB | 14.9 GB | 15 GB | noticeable loss of detail | unsloth/LTX-2-GGUF → |
| Q2_K | 8.1 GB | 12.9 GB | 13 GB | heavy loss; a last resort | unsloth/LTX-2-GGUF → |
“Needed” = file + 4 GB working memory + 0.8 GB system reserve.
02By amount of VRAM
03Best GPU for LTX-2
The cheapest cards (by launch price) that run it well, and every card sorted by memory: best GPU for LTX-2 → Planning bigger images or longer clips? Open the calculator →
04By graphics card
| GPU | VRAM | Verdict | Best file | Needed |
|---|---|---|---|---|
| Desktop graphics cards | ||||
| RTX 2060 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 3050 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 2070 Super 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 2080 Super 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3050 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3060 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3060 Ti 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3070 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3070 Ti 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 4060 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 4060 Ti 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5050 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5060 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5060 Ti 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 7600 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 9050 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 9060 XT 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| Arc B570 10 GB | 10 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3080 10 GB | 10 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 2080 Ti 11 GB | 11 GB | Offload only | Q2_K | 12.9 GB |
| Arc B580 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 2060 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 3060 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 3080 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 3080 Ti 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 4070 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 4070 Super 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 4070 Ti 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 5070 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RX 7700 XT 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RX 9070 GRE 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| Arc A770 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 4060 Ti 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 4070 Ti Super 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 4080 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 4080 Super 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 5060 Ti 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 5070 Ti 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 5080 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 7600 XT 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 7800 XT 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 7900 GRE 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 9060 XT 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 9070 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 9070 XT 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 7900 XT 20 GB | 20 GB | Runs | Q5_K_M | 19.1 GB |
| Arc Pro B60 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| RTX 3090 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| RTX 3090 Ti 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| RTX 4090 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| RX 7900 XTX 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| Arc Pro B70 32 GB | 32 GB | Runs well | Q8_0 | 25.2 GB |
| RTX 5090 32 GB | 32 GB | Runs well | Q8_0 | 25.2 GB |
| Laptop GPUs | ||||
| RTX 3050 Laptop 4 GB | 4 GB | Not practical | Q2_K | 12.9 GB |
| RTX 3050 Ti Laptop 4 GB | 4 GB | Not practical | Q2_K | 12.9 GB |
| RTX 2060 Laptop 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 3050 Laptop 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 3060 Laptop 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 4050 Laptop 6 GB | 6 GB | Not practical | Q2_K | 12.9 GB |
| RTX 2070 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 2070 Super Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 2080 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 2080 Super Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3070 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3070 Ti Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 3080 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 4060 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 4070 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5050 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5060 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 5070 Laptop 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 7600M 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 7600M XT 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 7600S 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RX 7700S 8 GB | 8 GB | Offload only | Q4_K_M | 17.6 GB |
| RTX 4080 Laptop 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 5070 Laptop 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 5070 Ti Laptop 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RX 7800M 12 GB | 12 GB | Offload only | Q2_K | 12.9 GB |
| RTX 3080 Laptop 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 3080 Ti Laptop 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 4090 Laptop 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 5080 Laptop 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RX 7900M 16 GB | 16 GB | Tight | Q3_K_M | 14.9 GB |
| RTX 5090 Laptop 24 GB | 24 GB | Runs | Q6_K | 20.8 GB |
| Unified memory | ||||
| Radeon 8060S (Strix Halo) 96 GB | 96 GB | Runs well | 16-bit | 42.6 GB |
| Radeon 8065S (Gorgon Halo) 160 GB | 160 GB | Runs well | 16-bit | 42.6 GB |
| Apple Silicon Macs (by memory) | ||||
| Mac 16 GB | 12.7 GB | Offload only | Q2_K | 12.9 GB |
| Mac 18 GB | 14.4 GB | Tight | Q2_K | 12.9 GB |
| Mac 24 GB | 19.6 GB | Runs | Q5_K_M | 19.1 GB |
| Mac 32 GB | 26.8 GB | Runs well | Q8_0 | 25.2 GB |
| Mac 36 GB | 30.2 GB | Runs well | Q8_0 | 25.2 GB |
| Mac 48 GB | 40.2 GB | Runs well | Q8_0 | 25.2 GB |
| Mac 64 GB | 55.7 GB | Runs well | 16-bit | 42.6 GB |
| Mac 96 GB | 85 GB | Runs well | 16-bit | 42.6 GB |
| Mac 128 GB | 115.4 GB | Runs well | 16-bit | 42.6 GB |
| Mac 192 GB | 175.4 GB | Runs well | 16-bit | 42.6 GB |
| Mac 256 GB | 236.9 GB | Runs well | 16-bit | 42.6 GB |
| Mac 512 GB | 498.1 GB | Runs well | 16-bit | 42.6 GB |
05Text encoder, VAE and other files
| File | Folder | Size | When | |
|---|---|---|---|---|
| Gemma 3 12B FP4 mixed gemma_3_12B_it_fp4_mixed.safetensors | models/text_encoders | 9.4 GB | text encoder · smaller, recommended | Download → |
| Gemma 3 12B FP8 (scaled) gemma_3_12B_it_fp8_scaled.safetensors | models/text_encoders | 13.2 GB | text encoder · alternative | Download → |
| Gemma 3 12B BF16 gemma_3_12B_it.safetensors | models/text_encoders | 24.4 GB | text encoder · alternative | Download → |
| LTX-2 video VAE ltx-2-19b-dev_video_vae.safetensors | models/vae | 2.4 GB | with GGUF files · GGUF / diffusion-model-only route | Download → |
| LTX-2 audio VAE ltx-2-19b-dev_audio_vae.safetensors | models/vae | 0.2 GB | with GGUF files · GGUF / diffusion-model-only route | Download → |
| LTX-2 text embeddings connectors (projection) ltx-2-19b-dev_embeddings_connectors.safetensors | models/text_encoders | 2.9 GB | with GGUF files · GGUF / diffusion-model-only route | Download → |
| LTX-2 spatial upscaler x2 ltx-2-spatial-upscaler-x2-1.0.safetensors | models/latent_upscale_models | 1.0 GB | optional · official two-stage template | Download → |
The files the official ComfyUI workflows load next to the model. Sizes read from Hugging Face (2026-10-01).
06Where the files go in ComfyUI
| File | Folder | Loader node |
|---|---|---|
| Diffusion model (.safetensors: 16-bit, FP8, INT8) | ComfyUI/models/diffusion_models | Load Diffusion Model |
| GGUF file (.gguf) | ComfyUI/models/unet | Unet Loader (GGUF) - from the ComfyUI-GGUF node pack |
| Text encoder | ComfyUI/models/text_encoders | Load CLIP / DualCLIPLoader (or the GGUF versions) |
| VAE | ComfyUI/models/vae | Load VAE |
07Measured and reported results
| Label | GPU | Setup | Result | Peak VRAM | Date | Source |
|---|---|---|---|---|---|---|
| reported | Radeon 8060S (Strix Halo) 96 GB | LTX-2 BF16 (T2V) · 1280x720 “"workflow": "LTX2-T2V-BF16.json", ... "duration_seconds": 615.0017409324646” kyuz0 Strix Halo toolbox benchmark; cold run incl. model load; resolution/frames from benchmark page; I2V = 616.16 s; steps not stated | 615 s / clip (121 frames) | — | 2026-02-13 | raw.githubusercontent.com → |
| reported | RTX 5090 32 GB | LTX-2 19B NVFP4 · 720p “generating 143 frames at 720p takes about 66 seconds end‑to‑end in ComfyUI” Issue says this is ~30-40% slower than NVIDIA's expected 40-45 s; steps not stated | 66 s / clip (143 frames) | — | 2026-01-08 | github.com → |
Other people's numbers are copied as published, with a link. Send yours.
08Training a LoRA for LTX-2
| Trainer | VRAM | Type | Settings and quote | Source |
|---|---|---|---|---|
| SimpleTuner | 13 GB | example run | Field report: RamTorch offload incl. text encoder, 480p, 17 frames, batch 2, AMD 7900XTX; int8 no offload ~29-30 GB; bf16 no offload ~48 GB “RamTorch (incl. text encoder): ~13 GB VRAM used on an AMD 7900XTX.” | github.com → |
| diffusion-pipe | 24 GB | stated minimum | Only LTX-2.3 supported (T2I/T2V, no audio); blocks_to_swap 46 (max), low resolution, short video, low rank; 'might barely fit' “with that, it might barely fit in 24GB VRAM if the resolution, video length, and LoRA rank are all low enough.” | github.com → |
| LTX-2 trainer (official, Lightricks) | 32 GB | stated minimum | t2v_lora_low_vram.yaml: INT8 (quanto) base, 8-bit text encoder, AdamW8bit, rank 16, batch 1, gradient checkpointing; 80GB+ for standard config “For GPUs with 32GB VRAM (e.g., RTX 5090), use the low VRAM config” | github.com → |
reported From each trainer's own docs, 2026-09-25. Why the figures differ →
09Every file tracked for LTX-2
| File | Type | Size | Repo |
|---|---|---|---|
| ltx-2-19b-dev.safetensors (dev) | 16-bit | 43.29 GB | Lightricks/LTX-2 → |
| ltx-2-19b-distilled.safetensors (distilled) | 16-bit | 43.29 GB | Lightricks/LTX-2 → |
| ltx-2-19b-distilled-fp8.safetensors (distilled) | FP8 | 27.08 GB | Lightricks/LTX-2 → |
| ltx-2-19b-dev-fp8.safetensors (dev) | FP8 | 27.08 GB | Lightricks/LTX-2 → |
| ltx-2-19b-dev-fp4.safetensors (dev) | FP4 | 19.99 GB | Lightricks/LTX-2 → |
| ltx-2-19b-dev-BF16.gguf | BF16 | 37.77 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-F16.gguf | F16 | 37.77 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q8_0.gguf | Q8_0 | 20.41 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q6_K.gguf | Q6_K | 15.97 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q5_1.gguf | Q5_1 | 14.62 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q5_K_M.gguf | Q5_K_M | 14.34 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q5_0.gguf | Q5_0 | 13.66 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q5_K_S.gguf | Q5_K_S | 13.62 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q4_K_M.gguf | Q4_K_M | 12.84 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q4_1.gguf | Q4_1 | 12.30 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q4_K_S.gguf | Q4_K_S | 11.86 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q4_0.gguf | Q4_0 | 11.35 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q3_K_L.gguf | Q3_K_L | 10.72 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q3_K_M.gguf | Q3_K_M | 10.12 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q3_K_S.gguf | Q3_K_S | 9.47 GB | unsloth/LTX-2-GGUF → |
| ltx-2-19b-dev-Q2_K.gguf | Q2_K | 8.10 GB | unsloth/LTX-2-GGUF → |
Open weights (Jan 2026). Joint audio+video generation. Text encoder Gemma 3 12B (not fetched). GGUF transformer-only needs separate VAE/audio-VAE files (unsloth repo has vae/ and text_encoders/ folders).