Skip to main content

xAI's Imagine Video 1.5 Gets Smarter: New Reference Tools and 1080p

xAI has just given its video generation model, Imagine Video 1.5, a serious upgrade. The new version brings a suite of features that push AI video creation closer to professional filmmaking: image reference, voice reference, text-to-video generation, and native 1080p output. If you've been wrestling with keeping characters consistent across shots or controlling scenes precisely, this update might be your new best friend.

Right now, the text-to-video and 1080p generation are live on the Grok Imagine web version, iOS, and Android apps. The image and voice reference features are initially rolling out to SuperGrok Heavy and SuperGrok Plus users in the United States, with a broader release planned later. So, if you're not in that group yet, hang tight—it's coming.

So, what does this mean for creators? Imagine typing a description and getting a video, or uploading a reference image to lock in a character's look, a product's design, or a specific scene. Even cooler, you can combine a character image with a voice sample, and the AI will keep the same face and voice across different shots. xAI says Imagine Video 1.5 can handle up to seven visual references per generation, which helps nail down facial features, product details, or environmental elements. That's a lot of control.

In multi-reference mode, you can swap backgrounds while keeping the character intact, change the character while preserving the scene, or tweak just the action while everything else stays put. The voice reference feature tackles a common headache in AI video: keeping a character's voice consistent. Now, the same character can sound the same across multiple segments, which is a game-changer for storytelling.

For developers, the xAI API now supports image references, text-to-video, and native 1080p generation through the grok-imagine-video-1.5 model. You can pass parameters like prompts, reference image URLs, video duration, aspect ratio, and resolution. The voice reference feature requires integration via request methods, so there's a bit of technical setup, but the capability is there.

Imagine Video 1.5 originally launched last month, and xAI had already optimized motion, physics simulation, and audio generation. This upgrade takes it further, enhancing control over identity, scenes, and product details. It's a clear sign that AI video is moving from simple image transitions to a multimodal tool that feels closer to a professional production pipeline.

Whether you're a content creator, a marketer, or just someone curious about AI's creative potential, this update is worth exploring. The ability to maintain consistency and control at 1080p resolution opens up new possibilities for everything from short films to product demos. And with the API access, developers can integrate these features into their own apps, making AI video generation more accessible than ever.

Key Points

  • New Features: Image reference, voice reference, text-to-video, and native 1080p output.
  • Availability: Text-to-video and 1080p on Grok Imagine apps; reference features for SuperGrok users in the US.
  • Multi-Reference: Up to seven visual references for precise control over characters, scenes, and products.
  • Voice Consistency: Combines character images with voice samples for stable voice performance.
  • Developer Access: API supports image references, text-to-video, and 1080p generation via grok-imagine-video-1.5.