LTX-2.5

All projects

Apps & Tools

Open world model for generating synchronized high-fidelity video and audio

byLightricks

Playable

LTX-2.5 media is blocked

Allow external media to connect to the provider and play this content.

Description

LTX-2.5 is an open-weights multimodal foundation model by Lightricks that generates synchronized video and audio from text, image, or video prompts. It features native multishot generation for scene consistency across cuts, dynamic compute allocation via 'diffusion fidelity rendering', and a custom Gemma 4 text encoder for complex prompt adherence.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: A major open-weights release that generates synchronized A/V and supports multishot scenes with consistent characters…
Attributed to: @Lightricks
Built with: Seedance 2.5 GPT-5.5 Pro Gemma 4 E2B/E4B LTX 2.3 Pro 3.1 Pro Vidu Q3 Pro Seedream 5.0 Pro Seedance 1.5 Pro Flux 2 Pro Nano Banana Pro Gemini 3 Pro Redwood AI Gemini 2.5 Pro
Links: huggingface.co
View tweet
Tags: #video-generation #audio-generation #open-weights #multimodal

Discovered: X
First posted: X · Aug 11, 2026
Listed: Added 9d ago
Build: Strong
Models: Strong

Prompt

A golden retriever running through a sunny meadow, cinematic lighting

Source

LTX-2.5

Description

Timeline

LTX-2.5

Description

Timeline