13.04B dev weights with base tag
Captured metadata reports 13,042,569,344 parameters, and hub tags record a base-model reference to Lightricks/LTX-Video with the LTXConditionPipeline pipeline tag.
Open Source Model Profile · Lightricks
LTX-Video-0.9.7-dev
LTX-Video-0.9.7-dev is a 13.04B-parameter text-to-video dev checkpoint from Lightricks. According to the model card, it supports text-to-video and image-plus-text-to-video generation through Diffusers LTXConditionPipeline.
- Publisher
- Lightricks
- Task
- text-to-video
- Model type
- Unknown
- License
- other
- Library
- diffusers
- Publication status
- Accepted · not indexed
Model overview
LTX-Video-0.9.7-dev is published by Lightricks as a text-to-video checkpoint. Captured Safetensors metadata reports 13,042,569,344 parameters, and hub tags record the Diffusers library with ltx-video, image-to-video, and a base-model reference to Lightricks/LTX-Video. According to the model card, this card focuses on the LTX-Video 0.9.7 model family with inference code in the linked codebase.
Recorded capabilities
Text-to-video and image-to-video
According to the model card, the model serves both text-to-video and image-plus-text-to-video use cases.
Diffusers compatibility
According to the model card, LTX Video works with the Diffusers library via LTXConditionPipeline and LTXLatentUpsamplePipeline, including from_single_file loading from original checkpoints.
Resolution and prompting guidance
According to the model card, resolutions should be divisible by 32 with frame counts divisible by 8 plus 1, and prompts should be elaborate English since prompt following depends on prompting style.
Use cases in the source record
- Text-to-video generation from elaborate English prompts using inference.py or the Diffusers LTXConditionPipeline workflow, with resolutions divisible by 32 and frame counts of 8 plus 1.
- Image-to-video generation from an input image plus prompt, with the model card documenting LTXVideoCondition conditioning and a latent-upsample plus denoise workflow.
Limitations and unknowns
- The captured license value is other, and the model card points to license-scoped direct use; exact license terms were not extracted.
- According to the model card, the highest-quality 13b dev variant requires more VRAM and distilled variants trade slight quality for speed, but no exact hardware requirement for this file was extracted.
- No evaluation results were extracted; real-time and quality descriptions are publisher claims only.
- Provider state is historical snapshot data, not independently refreshed current availability.
Source and provenance
Source: Lightricks/LTX-Video-0.9.7-dev
Captured: Unknown. Processed: 2026-09-07T19:35:12.877634+00:00.
LTX-Video 0.9.7 Model Card This model card focuses on the model associated with the LTX-Video model, codebase available here . LTX-Video is the first DiT-based video generation model capable of generating high-quality videos in real-time. It produces 30 FPS videos at a 1216×704 resolution faster than they can be watched. Trained on a large-scale dataset of diverse videos, the model generates high-resolution videos with realistic and varied content. We provide a model for both text-to-video as well as image+text-to-video usecases A woman with long brown hair and light skin smiles at another woman... A woman with long brown hair and l…
F001F002F003F004F005F006F007F008F009F010F011F014F015F016F017F019F021F027F028F029F035F036