Grok Imagine 1.5 is xAI's next-generation AI video and image creation model. Released on May 31, 2026, it builds on Grok Imagine 1.0 with better video quality, stronger prompt following, improved audio and faster performance. According to Arena.ai testing, the preview version gained about 52 Elo points compared to the earlier model.
One of the biggest updates is audio quality and how it matches actions and events in generated videos.
The model improves video extension features. Longer video sequences keep character appearance, movement, lighting and scene details more consistent. This helps reduce quality drops when extending clips multiple times.
Face quality has also improved. Testing showed better facial detail and stronger character consistency across frames. The model can create more realistic-looking people and maintain their appearance through longer scenes.
Photorealism got a major upgrade with videos showing sharper details, more natural lighting.
Prompt following is stronger as well. Grok Imagine 1.5 handles detailed instructions and complex scenes more accurately. Users can describe multiple elements and expect results that stay closer to the original request.
xAI also improved speed and efficiency. The model generates content faster, runs more reliably and offers broader support for reference images. These updates aim to make the video creation process smoother and quicker.
The preview version of Grok Imagine Video 1.5 reportedly reached the top position on the Image-to-Video Arena leaderboard. The model ranked ahead of leading competitors from ByteDance, Alibaba and Google, while showing a notable score increase over Grok Imagine 1.0.
