LTX Video-2.5 Extreme Uncensored Turbo v1.1 (Eros10 NSFW & 4-step LoRA baked-in)
UPDATED 9/10/2026:
- The latest LTX Video-2.5 Uncensored Turbo v1.1.
fp8_scaledfor speed. - Uncensored Gemma4 12B text encoder w/ LTX projections baked-in (no extra files needed).
int8for speed and accuracy.
Modes: img2video, txt2video, FL2VA / T2VA / I2VA / REF2VA/ AUDIO-TO-VIDEO.
New in v1.1:
A pre-merged, GGUF and FP8 build of LTX Video-2.5 with five fine-tunes baked into the weights:
- Eros10 NSFW LoRA - Eros10 is known for high quality, coherence, and Extreme NSFW (0.7 strength)
- DMD Distilled LoRA - DMD distilled LoRA promises faster generations, better facial preservation (I2V), and MUCH better instruction following than v1 or stock ltx25. (1.0 strength)
- LTX Video IC-LoRA - Official LTXV In-Context LoRA for better reference image adherence. (0.7 strength)
- LTX Video IC-LoRA Detailer - Allows for video-to-video control on top of text-to-video and image-to-video (0.45 strength)
- Img2Vid Lora - The BEST Img2Vid adapter for LTX-2, for better coherence and instruction following and easy I2V. (0.45 strength)
IC-LoRA Template Formatting (optional)
The IC LoRA is baked-in, use the template format for ultimate precision.
- Note: When invoking the IC-LoRA, the prompt must be formatted with the following tags:
[VISUAL]
[SPEECH]
[SOUNDS]
- Visual - Describe what is seen, and camera movement. LTX 2.5 supports multi-shots (ie. "start off front view close-up, then cut to a rear wide-view")
- Speech - Describe what is being said. Use
[brackets]for audio cues during speech (ie. [screams] [laughs] [sighs] [grunts]) - Sounds - Describe what is audible or not audible (the model ignores "no music" so use "unscored" to enforce no background music.)
Samples - Text-to-Video (4-steps)
Text-to-Video works with as few as 4-steps, use 8-steps for the best results.
a V12 supercar with neon lights underbody, speeding on a wet road. engine revs, turbo shifts, unscored b roll footage
Text-to-Video IC LoRA Prompt Template
Use the workflow Link
- Uses
comfyui-gguf-loadercustom_node
Samples - Image-to-Video
Image-to-Video is easy with this LTX 2.5 fine-tune (Img2Vid Adapter baked-in), just supply the image and the script, and the model handles the rest.
- Note: use 8-12 steps for best results.
- Higher CFG = more action, more audio, more talking.
Image-to-Video IC LoRA Prompt Template
Use the IC LoRA prompt format to control the action:
Use the workflow Link
- Uses
comfyui-gguf-loadercustom_node - Uses
RTX Super Resolutioncustom_node
Recommended settings
| Parameter | Value | Notes |
|---|---|---|
width × height | 1280 × 720 | 720P max resolution |
num_frames | 124 | 960 max frames |
frame_rate | 24.0 | |
steps | 8 | 8-steps minimum, 16 steps max |
guidance / CFG | 1.0 | between 1.0 - 4.0 for best results |
negative_prompt | (empty in samples) | optional |
Components used to generate
| Component | File | ~Size | Download |
|---|---|---|---|
| Transformer | ltx25_uncensored_v1.1-fp8_scaled.safetensors | 21.5 GB | Link |
| Text Encoder | gemma4_12b_ltx25_uncensored-int8.safetensors | 13.2 GB | Link |
| Audio VAE | ltx25_uncensored_v1.1_audio_vae.safetensors | 365 MB | Link |
| Video VAE | ltx25_uncensored_v1.1_video_vae.safetensors | 1.47 GB | Link |
- Note: These are ALL NEW files, NOT Sulphur
⚠ Must use the updated GGUF Loader ⚠
In comfy:
- Open the ComfyUI Manager
- Change the channel to "Channel (remote)"
- and search for comfyui-gguf-loader
- download latest build (nightly)
