Grok Imagine Video 1.5 Lite reaches 1080p by upscaling 720p

The price list has not moved. xAI's rate card for grok-imagine-video-1.5-lite still reads $0.02 a second at 480p, $0.03 at 720p and $0.14 at 1080p, the same three numbers it carried when the model surfaced through fal. Three later reads change what those numbers mean for anyone picking a video model this week.

The first is the top of the range. xAI's generation guide states that grok-imagine-video-1.5-lite "reaches 1080p for text-to-video and image-to-video by upscaling 720p". The standard grok-imagine-video-1.5 renders 1080p natively. Lite's $0.14-a-second 1080p rate buys an upscale of a 720p render, not a native 1080p pass.

What a minute of Lite costs

Artificial Analysis, read on October 16, 2026, prices a minute of 1080p video on each maker's own API at the model's default settings. The board puts Lite at $8.40 a minute, above Veo 3.1 Fast at $7.20, MiniMax H3 at $4.80, Vidu Q3 Turbo at $3.90, Boreal at $2.69 and Agnes-Video-2.5 at $2.40, and below grok-imagine-video-1.5 at $15.00, Wan 3.0 at $12.00 and Gemini Omni Flash 1.1 at $9.12. Lite's $8.40 is three and a half times Agnes-Video-2.5's $2.40. xAI's own ladder on the pricing page divides the same way: $0.02, $0.03 and $0.14 a second are $1.20, $1.80 and $8.40 a minute, so the two-cent figure holds at 480p and stops holding at 1080p.

The third read is about sound. xAI's video overview credits Lite with native audio and lip-synced speech. Treza Labs, an integrator that runs the model inside its own pipeline, wrote that clips came back with an audio track that was close to silent in its renders, far below normal listening level, and tells readers to treat the model as picture. The maker describes a capability, the integrator a measured level; the accounts disagree, so test your own prompt before a talking-head shot goes to Lite.

Lite is not only an xAI endpoint. It runs on fal, whose text-to-video API and image-to-video endpoint price output per second by resolution exactly as xAI's docs do, $0.02, $0.03 and $0.14, plus $0.01 for each input image, and puts a five-second clip with one input image at $0.11 at 480p, $0.16 at 720p and $0.71 at 1080p. fal's full-model page runs the flagship one tier higher, at $0.08, $0.14 and $0.25 a second, with sound generated in the same pass. Venice has listed Lite since September 30, 2026, priced from $0.04 a clip, with 1 to 15 second clips at up to 1080p and native audio.

Against a rival, Kuaishou's Kling pricing page lists Kling 3.0 Turbo with audio at 0.8 yuan a second at 720p and 1.0 yuan a second at 1080p, roughly $0.11 and $0.14. Lite's 720p rate of $0.03 sits below Kling's $0.11, and its $0.14 at 1080p lands level with Kling's.

For a workflow, the rate card and the upscale note point the same way. Draft on Lite at 480p for two cents a second, where a ten-second clip costs 20 cents, then finish the keeper on a model that renders 1080p natively, because Lite's 1080p rate runs seven times its 480p rate for an upscaled frame. The number that decides it is the effective cost per kept clip, which no price list publishes and retries do not leave out.

Grok Imagine Video 1.5 Lite is live on xAI's Imagine API as grok-imagine-video-1.5-lite, in the us-east-1 and us-west-2 regions, and through fal and Venice; set your resolution on every call, because the default is 480p.