Lightricks has released LTX-2.5, an open-weights model designed for local execution and fine-tuning. This new version enhances video and audio generation from various inputs, featuring native multishot generation for connected scenes and improved diffusion fidelity rendering. LTX-2.5 also includes a new diffusion video decoder, a custom Gemma 4-12B text encoder for complex prompts, and a prompt enhancer to expand instructions. AI
IMPACT Enhances capabilities for generating synchronized video and audio, with potential applications in robotics and physical AI.
RANK_REASON Model release from a recognized AI lab. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Hugging Face Trending Models →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →