base_model:

Lightricks/LTX-2.5 base_model_relation: quantized language:
en
de
es
fr
ja
ko
zh
it
pt library_name: gguf license: other license_name: ltx-2.x-community-license-agreement license_link: https://github.com/Lightricks/LTX-2/blob/main/LICENSE-2_x pipeline_tag: image-to-video arxiv: 2601.03233 tags:
gguf
unsloth
comfyui
text-to-video
image-to-video
text-to-audio-video
image-to-audio-video
ltx-2
ltx-2.5
ltx-video
ltxv
lightricks widget:
text: 'A blue humanoid robot with a glowing yellow ring for a face stands in a bright warehouse aisle (UD-Q4_K_XL)' output: url: samples/ltx25_distilled_UD-Q4_K_XL_t2v.mp4

This is a GGUF quantized version of LTX-2.5 by Lightricks, the 22B audio-video DiT, both the distilled (8 + 3 steps, CFG 1) and the dev (full, 30 steps, CFG 3) checkpoint. unsloth/LTX-2.5-GGUF uses Unsloth Dynamic 2.0 methodology: the UD- files are mixed precision, with every layer's type chosen from a measured sensitivity scan, and they beat the uniform files at the same size.

Loads in Unsloth Studio and in ComfyUI through ComfyUI-GGUF (Unet Loader (GGUF)), both tested on every file here.
Same layout as unsloth/LTX-2.3-GGUF: dev files at the root, distilled files in distilled/. The text encoder, VAEs and latent upscaler are unchanged; take them from Lightricks/LTX-2.5.

Recommended: UD-Q4_K_XL (14.0 GB) for most GPUs: clearly better than Q4_K_M and on par with Q5_K_M, which is 1.6 GB larger. For output closest to bf16 use Q8_0; on small cards use UD-Q2_K_XL (9.5 GB), which is far cleaner than Q2_K and close to Q3_K_M at 1.8 GB less.

Samples

distilled/ltx-2.5-22b-distilled-UD-Q4_K_XL.gguf, 1280x704, 121 frames, 24 fps, seed 42, with audio: text to video, image to video.

Frame 60 of the same text-to-video render from each file:

And of the image to video render:

Files and accuracy

LPIPS (lower is better) and PSNR against the bf16 DiT, same prompt, seed, text encoder and settings, mean over every one of the 121 frames. Distilled: 1280x704, 121 frames, 8 + 3 two-stage steps, CFG 1, text to video and image to video, 4 seeds (8 clips). Dev: 1280x704, 121 frames, 30 steps, CFG 3, text to video and image to video, 2 seeds (4 clips). Both run deterministically, so a repeated bf16 render scores 0 and the numbers measure only the quantization. The samplers re-inject noise every step, so small weight errors move the motion more than the look of a frame: every file keeps the subject, scene and prompt.

For scale: the hosted unsloth/LTX-2.5-FP8 INT8-ConvRot and FP8 DiTs score 0.265 and 0.366 on the same clips in Studio's default (compiled) mode, where compiled versus uncompiled bf16 alone already differs by 0.265.

Distilled (distilled/):

File	Size	Bits per weight	LPIPS vs bf16	PSNR vs bf16
`distilled/ltx-2.5-22b-distilled-Q2_K.gguf`	8.46 GB	3.22	0.567	14.8 dB
`distilled/ltx-2.5-22b-distilled-UD-Q2_K_XL.gguf`	9.50 GB	3.62	0.416	17.1 dB
`distilled/ltx-2.5-22b-distilled-Q3_K_M.gguf`	11.30 GB	4.30	0.390	17.3 dB
`distilled/ltx-2.5-22b-distilled-UD-Q3_K_XL.gguf`	12.00 GB	4.57	0.310	19.1 dB
`distilled/ltx-2.5-22b-distilled-Q4_K_M.gguf`	13.58 GB	5.17	0.352	18.1 dB
`distilled/ltx-2.5-22b-distilled-UD-Q4_K_XL.gguf`	14.00 GB	5.33	0.320	19.4 dB
`distilled/ltx-2.5-22b-distilled-Q5_K_M.gguf`	15.63 GB	5.95	0.316	18.9 dB
`distilled/ltx-2.5-22b-distilled-UD-Q5_K_XL.gguf`	16.00 GB	6.09	0.296	20.3 dB
`distilled/ltx-2.5-22b-distilled-Q6_K.gguf`	17.81 GB	6.78	0.274	20.4 dB
`distilled/ltx-2.5-22b-distilled-Q8_0.gguf`	22.78 GB	8.68	0.242	21.6 dB
`distilled/ltx-2.5-22b-distilled-BF16.gguf`	42.03 GB	16.01	0 (bit-identical)	lossless

Dev (repo root):

File	Size	Bits per weight	LPIPS vs bf16	PSNR vs bf16
`ltx-2.5-22b-dev-Q2_K.gguf`	8.46 GB	3.22	0.592	14.3 dB
`ltx-2.5-22b-dev-UD-Q2_K_XL.gguf`	9.50 GB	3.62	0.354	16.6 dB
`ltx-2.5-22b-dev-Q3_K_M.gguf`	11.30 GB	4.30	0.427	15.4 dB
`ltx-2.5-22b-dev-UD-Q3_K_XL.gguf`	12.00 GB	4.57	0.246	20.0 dB
`ltx-2.5-22b-dev-Q4_K_M.gguf`	13.58 GB	5.17	0.330	16.8 dB
`ltx-2.5-22b-dev-UD-Q4_K_XL.gguf`	14.00 GB	5.33	0.183	21.4 dB
`ltx-2.5-22b-dev-Q5_K_M.gguf`	15.63 GB	5.95	0.208	20.1 dB
`ltx-2.5-22b-dev-UD-Q5_K_XL.gguf`	16.00 GB	6.09	0.162	22.1 dB
`ltx-2.5-22b-dev-Q6_K.gguf`	17.81 GB	6.78	0.180	21.2 dB
`ltx-2.5-22b-dev-Q8_0.gguf`	22.78 GB	8.68	0.169	23.4 dB
`ltx-2.5-22b-dev-BF16.gguf`	42.03 GB	16.01	reference	reference

BF16 is a lossless repack: rendered through the GGUF path it is bit-identical to the original bf16 safetensors.

Run in Unsloth Studio

Pick LTX-2.5 on the Video page and choose one of these GGUF files. Studio loads the DiT from the GGUF and the text encoder, VAEs, connectors and upsampler from Lightricks/LTX-2.5-Diffusers, and runs the distilled files with the two-stage 8 + 3 step recipe and the dev files at 30 steps, CFG 3.

Run in ComfyUI

Use an up to date ComfyUI (with LTX-2.5 support) and the ComfyUI-GGUF custom node.

cd ComfyUI/custom_nodes
git clone https://github.com/city96/ComfyUI-GGUF.git
pip install -r ComfyUI-GGUF/requirements.txt
cd ../models
ln -s "$(hf download unsloth/LTX-2.5-GGUF distilled/ltx-2.5-22b-distilled-UD-Q4_K_XL.gguf --quiet)" unet/.
ln -s "$(hf download Lightricks/LTX-2.5 text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors --quiet)" text_encoders/.
ln -s "$(hf download Lightricks/LTX-2.5 vae/ltx-2.5-video-vae-bf16.safetensors --quiet)" vae/.
ln -s "$(hf download Lightricks/LTX-2.5 vae/ltx-2.5-audio-vae-bf16.safetensors --quiet)" vae/.
ln -s "$(hf download Lightricks/LTX-2.5 latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors --quiet)" latent_upscale_models/.

Open ComfyUI's built-in Text to Video (LTX-2.5) template and replace its Load Diffusion Model node with Unet Loader (GGUF) pointing at the GGUF file. Everything else in the template stays as it is.

Licence

LTX-2.x Community License Agreement, from Lightricks/LTX-2.5. Full text in LICENSE. These files are quantized derivatives of the original weights; not an official Lightricks release.

LTX-2.5 is an open world model with open weights, built for local execution and fine-tuning. Its established use is generating synchronized, high-fidelity video and audio from text, image, and video inputs; applicability to emerging domains such as robotics and physical AI is developing.

Full control and customization — self-host on your own infrastructure. No per-generation billing, no per-seat lock-in, no forced API dependency. Revenue is measured across the whole entity, including subsidiaries and affiliates under common control. The full, binding terms live in LICENSE.

What's new in LTX-2.5

Native multishot generation — generate connected scenes in a single pass: multiple shots that hold character identity, environment, lighting, voice, and visual style across cuts (previous versions produced a single continuous shot).
Diffusion fidelity rendering — Instead of locking every scene to one compression rate, our model dynamically allocates compute by scene complexity and budget, rendering flawless detail where it matters, efficient everywhere else.
New diffusion video decoder — replaces the VAE reconstruction stage; sharper faces, textures, and on-screen text, better motion, and fewer artifacts in demanding scenes.
Custom Gemma 4 12B text encoder — holds complex prompts together (multiple characters, camera moves, lighting, actions) instead of dropping details across a longer sequence.
Prompt enhancer — expands a short prompt into richer cinematic instructions at minimal extra compute.
Duration predictor (optional) — an opt-in node predicts a clip's length from the prompt and sets the frame count for you, instead of relying on a fixed-duration parameter.
Substantially improved distilled model — retains much more of the full model's visual quality, prompt adherence, and motion consistency in a smaller, faster checkpoint.

Model family & checkpoints

LTX-2.5 ships as a split, Comfy-aligned pack (one .safetensors per component) rather than a single monolith. Point each CLI flag / loader at the file below.

Transformers (DiT)

File	Notes
`diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors`	Distilled DiT (bf16). Fixed 8-step schedule, CFG=1.
`diffusion_models/ltx-2.5-22b-dev-transformer-bf16.safetensors`	Full / trainable DiT (bf16).
`diffusion_models/ltx-2.5-22b-distilled-transformer-comfy-int8-convrot.safetensors`	Distilled DiT (Comfy int8 + convrot). ComfyUI only — not for `ltx-pipelines` / PyTorch.
`diffusion_models/ltx-2.5-22b-dev-transformer-comfy-int8-convrot.safetensors`	Full DiT (Comfy int8 + convrot). ComfyUI only — not for `ltx-pipelines` / PyTorch.
`diffusion_models/ltx-2.5-22b-distilled-transformer-nvfp4.safetensors`	Distilled DiT (NVFP4). ComfyUI, or `ltx-pipelines` with `--quantization nvfp4-prequant` (Blackwell / `ltx-kernels`).

Other components

File	Notes
`text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors`	Gemma4 TE + projections (bf16)
`text_encoders/gemma4-12b-with-proj-ltx-2.5-comfy-int8-convrot.safetensors`	Same TE, Comfy int8 — ComfyUI only
`vae/ltx-2.5-video-vae-bf16.safetensors`	DiffVAE — higher quality, heavier
`vae/ltx-2.5-video-vae-conv-bf16.safetensors`	Conv VAE — faster, lighter
`vae/ltx-2.5-audio-vae-bf16.safetensors`	Audio VAE + vocoder
`loras/ltx-2.5-22b-distilled-lora-450-bf16.safetensors`	Distilled LoRA (dev-transformer workflows)
`model_patches/ltx-2.5-duration-head-bf16.safetensors`	Auto duration when `--num-frames` omitted
`latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors`	x2 spatial upscaler required for multi-stage pipeline
`latent_upscale_models/ltx-2.5-latent-temporal-upscaler-x2-bf16-1.0.safetensors`	x2 temporal upscaler

Usage

Online demo

Try LTX-2.5 in the API Playground without installing anything locally.

Option A — Python (`ltx-pipelines`)

Weights on this repo are split (Comfy-aligned): one safetensors file per component. The LTX-2 ltx-pipelines package loads them via --transformer-path, --text-encoder-path, etc.

Install

git clone https://github.com/Lightricks/LTX-2.git
cd LTX-2
uv sync
source .venv/bin/activate

Python >= 3.12, CUDA >= 12.7, PyTorch ~= 2.7 recommended. See the repo README for attention backends and optional extras.

Download weights

hf auth login

# LTX-2.5 distilled split pack
hf download Lightricks/LTX-2.5 \
  diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
  text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
  vae/ltx-2.5-video-vae-bf16.safetensors \
  vae/ltx-2.5-audio-vae-bf16.safetensors \
  model_patches/ltx-2.5-duration-head-bf16.safetensors \
  latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
  --local-dir models/ltx-2.5

Distilled text-to-video

uv run python -m ltx_pipelines.distilled \
  --transformer-path     models/ltx-2.5/diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
  --text-encoder-path    models/ltx-2.5/text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
  --video-vae-path       models/ltx-2.5/vae/ltx-2.5-video-vae-bf16.safetensors \
  --audio-vae-path       models/ltx-2.5/vae/ltx-2.5-audio-vae-bf16.safetensors \
  --duration-head-path   models/ltx-2.5/model_patches/ltx-2.5-duration-head-bf16.safetensors \
  --spatial-upsampler-path models/ltx-2.5/latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
  --prompt "A golden retriever running through a sunny meadow, cinematic lighting" \
  --seed 42 \
  --output-path output_distilled.mp4

Omit --num-frames to let the duration head pick a length from the prompt (LTX-2.5+). Or set e.g. --num-frames 121 (must satisfy frames % 8 == 1). Width/height must be divisible by 32.

Image-to-video

Add one or more --image PATH FRAME_IDX STRENGTH flags (frame 0 = first frame conditioning):

uv run python -m ltx_pipelines.distilled \
  --transformer-path     models/ltx-2.5/diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
  --text-encoder-path    models/ltx-2.5/text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
  --video-vae-path       models/ltx-2.5/vae/ltx-2.5-video-vae-bf16.safetensors \
  --audio-vae-path       models/ltx-2.5/vae/ltx-2.5-audio-vae-bf16.safetensors \
  --duration-head-path   models/ltx-2.5/model_patches/ltx-2.5-duration-head-bf16.safetensors \
  --spatial-upsampler-path models/ltx-2.5/latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
  --image path/to/first_frame.jpg 0 1.0 \
  --prompt "The camera slowly dollies out as wind moves through the grass" \
  --seed 42 \
  --output-path output_i2v.mp4

Low-VRAM tips

# Downcast bf16 transformer on the fly + CPU offload
  ...existing flags... \
  --quantization fp8-cast \
  --offload cpu

Use the bf16 checkpoints with ltx-pipelines. The *-comfy-int8-convrot.safetensors files are ComfyUI-only and are not loaded by this PyTorch path.

Python API (same split paths)

from ltx_pipelines.distilled import DistilledPipeline
from ltx_pipelines.utils.model_paths import ModelPaths

model_paths = ModelPaths.from_split(
    transformer_path="models/ltx-2.5/diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors",
    text_encoder_path="models/ltx-2.5/text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors",
    video_vae_path="models/ltx-2.5/vae/ltx-2.5-video-vae-bf16.safetensors",
    audio_vae_path="models/ltx-2.5/vae/ltx-2.5-audio-vae-bf16.safetensors",
    duration_head_path="models/ltx-2.5/model_patches/ltx-2.5-duration-head-bf16.safetensors",
)

pipe = DistilledPipeline(
    model_paths=model_paths,
    spatial_upsampler_path="models/ltx-2.5/latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors",
)
# See packages/ltx-pipelines for __call__ args (prompt, seed, num_frames, images, ...).

uv run python -m ltx_pipelines.distilled --help

Full docs: ltx-pipelines installation.

Option B — ComfyUI

Official LTX-2.5 workflow templates ship in ComfyUI. Full instructions: ComfyUI integration.

Option C — Diffusers

A Diffusers-compatible pack lives at Lightricks/LTX-2.5-Diffusers — same model, Diffusers-friendly packaging.

Install

LTX-2.5 support is not in a diffusers release yet, so install from main:

pip install git+https://github.com/huggingface/diffusers

Image-to-video, two stages

import torch
from diffusers import LTX2ImageToVideoPipeline, LTX2LatentUpsamplePipeline
from diffusers.pipelines.ltx2.latent_upsampler import LTX2LatentUpsamplerModel
from diffusers.pipelines.ltx2.utils import (
    DEFAULT_NEGATIVE_PROMPT,
    DISTILLED_SIGMA_VALUES,
    STAGE_2_DISTILLED_SIGMA_VALUES,
)
from diffusers.utils import encode_video, load_image

MODEL_ID = "Lightricks/LTX-2.5-Diffusers"
# Stage 1 resolution; stage 2 runs at 2x this.
HEIGHT, WIDTH, NUM_FRAMES, FRAME_RATE = 544, 960, 121, 24.0

pipe = LTX2ImageToVideoPipeline.from_pretrained(MODEL_ID, dtype=torch.bfloat16)
pipe.enable_model_cpu_offload()
pipe.vae.enable_tiling()  # stage 2 decodes at 2x

latent_upsampler = LTX2LatentUpsamplerModel.from_pretrained(
    MODEL_ID, subfolder="latent_upsampler", dtype=torch.bfloat16
).to("cuda")
upsample_pipe = LTX2LatentUpsamplePipeline(vae=pipe.vae, latent_upsampler=latent_upsampler)

generator = torch.Generator("cuda").manual_seed(42)
shared = dict(
    image=load_image("path/to/first_frame.jpg"),
    prompt="The camera slowly dollies out as wind moves through the grass",
    negative_prompt=DEFAULT_NEGATIVE_PROMPT,
    frame_rate=FRAME_RATE,
    guidance_scale=1.0,
    audio_guidance_scale=1.0,
    stg_scale=0.0,
    audio_stg_scale=0.0,
    modality_scale=1.0,
    audio_modality_scale=1.0,
    generator=generator,
    return_dict=False,
)

stage_1_latents, audio_latents = pipe(
    height=HEIGHT, width=WIDTH, num_frames=NUM_FRAMES,
    sigmas=DISTILLED_SIGMA_VALUES, output_type="latent", **shared,
)

upsampled_latents = upsample_pipe(
    latents=stage_1_latents, output_type="latent", return_dict=False
)[0]

# Stage 2 takes its size from the upsampled latents, so pass no height/width.
video, audio = pipe(
    num_frames=NUM_FRAMES,
    sigmas=STAGE_2_DISTILLED_SIGMA_VALUES,
    latents=upsampled_latents,
    audio_latents=audio_latents,
    noise_scale=STAGE_2_DISTILLED_SIGMA_VALUES[0],
    output_type="np",
    **shared,
)

encode_video(
    video[0],
    fps=int(FRAME_RATE),
    output_path="output_i2v_two_stage.mp4",
    audio=audio[0].float().cpu(),
    audio_sample_rate=pipe.vocoder.config.output_sampling_rate,
)

Constraints

Frame count: num_frames % 8 == 1 (1, 9, 17, …, 121, …)
Width and height divisible by 32

Prompting

Well-structured, detailed prompts materially improve results. For multishot prompting and a full guide, see How to prompt LTX-2.

Training & fine-tuning

The dev transformer is fully trainable. Reproduce published LoRAs and IC-LoRAs with the LTX-2 Trainer.

Based on our testing, the large majority of LoRAs and IC-LoRAs trained on LTX-2.3 run on LTX-2.5 without changes. A small number of exceptions exist — validate your adapters before production use.

Limitations

This model is not intended or able to provide factual information.
As a statistical model, this checkpoint may amplify existing societal biases.
Prompt following is heavily influenced by prompting style.
The model may fail to generate videos that match the prompt perfectly.
The model may generate content that is inappropriate or offensive.

Citation

@article{hacohen2025ltx2,
  title={LTX-2: Efficient Joint Audio-Visual Foundation Model},
  author={HaCohen, Yoav and Brazowski, Benny and Chiprut, Nisan and Bitterman, Yaki and Kvochko, Andrew and Berkowitz, Avishai and Shalem, Daniel and Lifschitz, Daphna and Moshe, Dudu and Porat, Eitan and Richardson, Eitan and Guy Shiran and Itay Chachy and Jonathan Chetboun and Michael Finkelson and Michael Kupchick and Nir Zabari and Nitzan Guetta and Noa Kotler and Ofir Bibi and Ori Gordon and Poriya Panet and Roi Benita and Shahar Armon and Victor Kulikov and Yaron Inger and Yonatan Shiftan and Zeev Melumian and Zeev Farbman},
  journal={arXiv preprint arXiv:2601.03233},
  year={2026}
}

LTX-2.5-GGUF

Samples

Files and accuracy

Run in Unsloth Studio

Run in ComfyUI

Licence

What's new in LTX-2.5

Model family & checkpoints

Transformers (DiT)

Other components

Usage

Online demo

Option A — Python (ltx-pipelines)

Install

Download weights

Distilled text-to-video

Image-to-video

Low-VRAM tips

Python API (same split paths)

Option B — ComfyUI

Option C — Diffusers

Install

Image-to-video, two stages

Constraints

Prompting

Training & fine-tuning

Limitations

Citation

Option A — Python (`ltx-pipelines`)