xAI Ships Native 1080p Video and Seven-Image Character References in Grok Imagine 1.5

xAI pushed a major update to Grok Imagine Video 1.5, adding native 1080p output and a Multi-Reference tool that accepts up to seven image references per generation. The upgrade applies to both text-to-video and image-to-video workflows, doubling the resolution ceiling from the model’s previous 720p cap.

The headline feature is Multi-Reference: creators can now preserve a character’s face across scene changes, keep a location fixed while swapping who’s in it, or hold both constant while only altering the action. Voice references are rolling out over the following days, extending the same consistency logic to audio — the same character face and voice can now persist across multiple generated shots, something short-form AI video tools have struggled with.

xAI describes Imagine Video 1.5, which launched last month, as its “best video model yet, with more lifelike motion and sound.” This update builds directly on that base rather than shipping a new model version.

Access starts narrow: SuperGrok Heavy and SuperGrok Plus subscribers in the US get first access via the Grok Imagine website, iOS app, and Android app. Text-to-video and 1080p generation are already generally available across all subscription tiers — it’s the Multi-Reference and voice-reference tools that are gated to the top plans for now. Developers can reach the same capabilities programmatically through the xAI API under the grok-imagine-video-1.5 model.

The move puts xAI in more direct competition with OpenAI’s Sora and Google’s Veo, both of which have pushed character-consistency tooling as a way to make AI video usable for actual production work rather than one-off clips. Multi-shot consistency has been the biggest practical blocker to using generative video for anything longer than a single scene — a director can specify a character once and reuse them, instead of re-prompting and hoping the model reproduces the same face.

xAI hasn’t published pricing changes alongside the update, and there’s no indication yet of when Multi-Reference and voice references reach lower subscription tiers or the free tier. Given how fast xAI has iterated on Imagine — this is already a follow-up release less than a month after Imagine Video 1.5’s launch — a broader rollout is likely to come quickly.

Sources

• xAI adds character references and 1080p to Imagine Video 1.5