xAI Upgrades Imagine Video 1.5: Adds Text-to-Video, Native 1080p, and Multi-References
testingcatalog · x · 2026-08-01
xAI officially announced a major upgrade to its video generation model, Imagine Video 1.5, introducing Text-to-Video and native 1080p resolution support. Users can now generate high-quality video directly from prompts without needing a starting image.
Additionally, the model introduces image and voice references, allowing users to lock in a specific character's face and voice across multiple scenes for high consistency. A single generation supports up to 7 reference images to lock characters, scenes, or products. Image and voice references are currently available to SuperGrok Heavy and Plus subscribers in the US, with API support also rolled out.
More from Multimodal
- UniFace: Open-Source Library Unifies Face Detection, Recognition, and Tracking — tom_doerr · 2026-08-01
- Training a Krea 2 LoRA for amateur smartphone-style photos with Indian vibe — desiNaughtyAI · 2026-08-01
- Experiment: 512-res generation + SeedVR2 upscaling matches 2048 output — More_Bid_2197 · 2026-08-01
- ByteDance uses Seedance 2.5 to push 'cinematic' lessons in study app — JayaIsNotGemez · 2026-08-01
- Midjourney Style Reference Code Powers Surreal AI Video Workflow — azed_ai · 2026-08-01
- voice-pro: A Gradio WebUI Integrating Zero-Shot Voice Cloning and Multilingual Translation — abus-aikorea · 2026-08-01