Grok Imagine Video 1.5 (Preview)

Spec

Attribute	Value
Developer	xai
Released	2026-06-03
Announced	2026-06-03
Context window	unknown
Pricing	$0.08 per second of output video
License	unknown
Availability	api.x.ai (preview as of June 3, 2026)
Model ID	grok-imagine-video-1.5-2026-05-30 (alias: grok-imagine-video-1.5-preview)
Task	Image-to-video generation
Max resolution	720p
Max clip length	15 seconds
Architecture	Aurora engine — autoregressive mixture-of-experts, interleaved text/image/video/audio modalities

Release Date

2026-06-03 (API preview)

Benchmarks / Rankings

• #1 on Artificial Analysis Video Arena Image-to-Video leaderboard (June 3, 2026)
• Faster inference than prior xAI video models

Use Cases

• Animating still images: provide a starting frame + motion prompt → cinematic video
• Camera moves, atmospheric effects, physics-consistent animation
• Social/marketing video production
• Developer integration via xAI API

Features

• Native audio sync (audio aligned to generated video)
• Faithful to source image (high source image adherence)
• Cinematic camera movement modeling
• Faster inference than grok-imagine-video-1.0

Compared To

Model	Provider	Max Res	Audio	Leaderboard
Grok Imagine Video 1.5	xAI	720p	✅ native	#1 Artificial Analysis (Jun 2026)
Gemini Omni	Google	high	✅	top 3
Sora 2	OpenAI	1080p	limited	top 5
Veo 3	Google	1080p	✅	top 3

Significance

• xAI's Aurora engine extends beyond text to unified multimodal token generation — the same architecture for voice, image, and video
• Complements Grok Imagine (still image) and Grok Voice; Aurora is becoming a full multimodal stack
• First xAI model to top the Video Arena leaderboard — validates the quality of the Aurora engine's video generation against specialized competitors

Sources

• Grok Imagine 1.5 Preview — xAI (snapshot)

Open Questions

• Is 1080p resolution planned?
• Text-to-video (not only image-to-video) via API?
• How will pricing evolve as the model exits preview?