Grok Imagine Video 1.5: Audio, References, 1080p and API Pricing

What Grok Imagine Video 1.5 is

Grok Imagine Video 1.5 is a video generation model from xAI, which now publishes as SpaceXSI. Its dated snapshot is grok-imagine-video-1.5-2026-05-30, and xAI announced it in the API in preview on 3 June 2026 as grok-imagine-video-1.5-preview, an image-to-video model that animates a single starting frame from a prompt describing the motion, at up to 720p. On 16 June 2026 xAI made it generally available in the API as grok-imagine-video-1.5 and rolled out a Video 1.5 Fast version on grok.com/imagine and the Grok apps for iOS and Android. On LMArena's video leaderboard Grok Imagine Video 1.5 has 1488 Elo (rank 5).

What changed from the previous model

xAI lists three improvements over its previous video model. Sound effects, ambience and dialogue are generated in the same pass as the picture and land on the action, with clearer and better-synced speech. Movement holds together across a clip, with fewer warps and more believable weight and momentum. And Video 1.5 Fast almost doubles generation speed: xAI states it produces a 6-second 720p clip in about 25 seconds, against more than 40 seconds for the previous model. The preview notes say the model keeps detail and lighting from the input frame, so a clip continues the original image rather than reinterpreting it.

References, text-to-video and 1080p

On 31 July 2026 xAI extended Video 1.5 with references. Image and voice references keep characters and voices consistent, starting in the US for SuperGrok Heavy and SuperGrok Plus subscribers on grok.com/imagine and iOS before rolling out to all tiers. Text-to-video removed the need for a starting image by pairing xAI's image generation with image-to-video, and native 1080p output became available for both text-to-video and image-to-video. xAI announced text-to-video and native 1080p as generally available on grok.com/imagine, iOS and Android.

Using it through the API

In the API a request gives a starting image, a prompt describing the motion, and the chosen resolution and duration; xAI's example asks for a 10-second clip at 720p. The model accepts text, images and audio and returns video, and grok-imagine-video-1.5-preview and the dated snapshot work as aliases of the same model. xAI bills each second of generated video by resolution: $0.08 at 480p, $0.14 at 720p and $0.25 at 1080p, so the 10-second 720p example costs $1.40. Each input image costs $0.01, and audio input using preset voices is free. The model page lists a rate limit of 10 requests per second in us-east-1, and the Batch API is supported in us-east-1 and us-west-2. A cheaper Grok Imagine Video 1.5 Lite is listed separately.

Workflow features in Grok Imagine

Alongside the model, xAI added features to the Grok Imagine apps: Projects to organise work in the sidebar, multiple agents that run several prompts in parallel instead of one after another, and search across every image and video a user has made. xAI also published a behind-the-scenes example of a film trailer made entirely with Grok Imagine 1.5.

Official Resources