Key Features

Open-source video foundation model.
Improved distilled generation model.
Cleaner motion from a new video decoder.
Native multishot scene consistency.
Prompt enhancement with a Gemma 4 12B encoder.
Automatic clip duration prediction.
Native 4K HDR and RAW workflows.
Fine-tunable base checkpoint and API access.

The release adds a better video decoder, a Gemma 4 12B text encoder with a custom prompt enhancer, beta precise video editing, native 4K HDR and RAW support, and a pretrained foundation checkpoint intended for fine-tuning. On two GB200 GPUs, LTX reports a 10-second 720p clip in 6.8 seconds.


LTX-2.5 supports both self-hosted experimentation and a paid API billed by generated video seconds. It is useful for creators, production teams, developers, and researchers who need controllable text-to-video, multishot generation, editing, or domain adaptation.

Get more likes & reach the top of search results by adding this button on your site!

Embed button preview - Light theme
Embed button preview - Dark theme
TurboType Banner

Subscribe to the AI Search Newsletter

Get top updates in AI to your inbox every weekend. It's free!