Grok Imagine Video 1.5 and Lite: release, price

Grok Imagine Video 1.5 is xAI's video model with native audio, in its API since June 2026: 1 to 15 seconds from text, a still or references, at up to native 1080p (720p for references). xAI lists $0.08 to $0.25 per second on 16 October 2026; 1.5 Lite starts at $0.02. On Versely, 1.5 starts at 7 credits.

Key facts

Item	Detail
Models	grok-imagine-video-1.5, the cheaper grok-imagine-video-1.5-lite and the earlier grok-imagine-video
Released	1.5: June 2026 in the API; a 31 July 2026 update added text-to-video, 1080p and reference-to-video. Lite: marked New in xAI's docs, no launch date
Resolution	480p, 720p or 1080p. Native on 1.5 except reference-to-video, which stops at 720p. Lite upscales 720p
Length	1 to 15 seconds, default 8
Audio	On by default, with lip-synced speech. generate_audio=false returns a silent clip
Aspect ratios	Nine, including 21:9 and 5:2. Image to video follows the still
References and pins	1.5: up to 14 images, 3 preset voices, first and last frame, up to 4 keyframes. Lite: none
xAI price per second, listed 16 October 2026	1.5: $0.08 at 480p, $0.14 at 720p, $0.25 at 1080p. Lite: $0.02, $0.03, $0.14
On Versely	Three 1.5 rows from 7 credits. No Lite row

Dates. xAI's docs give no launch day for either model. The 1.5 model page lists a dated alias, grok-imagine-video-1.5-2026-05-30; WinBuzzer (18 June 2026), citing an xAI post of 17 June, reported general availability. TechTimes dated the update that added text-to-video, native 1080p and references to 31 July; xAI's release notes list it under July. For Lite, the Video Overview that introduces it is stamped 8 October 2026 and LLM Gateway lists 8 October as the release, but LLMReference says 1 October and Artificial Analysis labels it Aug 2026. The Lite launch day is not settled, so treat 8 October as approximate.

Sources, all read on 16 October 2026: xAI's docs (titled "SpaceXSI Docs"): Video Overview, Video Generation, Reference-to-Video, release notes and the model pages for 1.5, Lite and the original; Artificial Analysis' text-to-video and image-to-video boards; and, to corroborate, WinBuzzer, TechTimes, TestingCatalog, LLM Gateway and LLMReference. xAI's news posts sit behind a bot check, so we did not read them.

What changed from the original

The original grok-imagine-video stays in xAI's lineup and alone edits and extends clips. xAI's Video Overview compares the three:

Capability	1.5	Lite	Original
Text and image to video	Up to 1080p, native	1080p, upscaled	Up to 720p
Reference images	Up to 14	None	Up to 7
Voice references	3 preset	None	None
Pinned frames	First/last, 4 keyframes	None	None
Edit or extend	No	No	Yes
Audio	Lip-synced speech	Lip-synced speech	Generated audio
Price at 480p, 720p	$0.08, $0.14	$0.02, $0.03	$0.05, $0.07

Sound. xAI lists 1.5 and Lite as native audio with lip-synced speech and the original as generated audio. On 1.5, reference-to-video can give up to three characters a preset voice. Voices from your own audio files are for trusted partners, on request.

Text to video. On 1.5 and Lite, xAI renders a first frame from your prompt, then animates it, in one request. When the opening frame matters, pick the still yourself and use image to video.

References. The Reference-to-Video page, stamped 9 October 2026, allows up to 14 images and 15 seconds on 1.5, against 7 images and 10 seconds on the original. TechTimes and TestingCatalog reported seven for the 31 July update; xAI's pages give no date for the change.

Pinned frames. Only 1.5 supports first-and-last-frame pins and up to four keyframes at chosen timestamps; the same still as first and last frame makes a loop.

How it ranks

On Artificial Analysis' blind-vote text-to-video board on 16 October 2026, 1.5 is 10th with an Elo of 1060 (rank range 9 to 12) and Lite is 17th at 999, while Wan 3.0 tops the board at 1174. On image to video, 1.5 is 8th at 1068, behind MiniMax H3 (768p) at 1141. Upper-middle on both, not the leader.

Video 1.5 or Lite?

xAI recommends Lite for simple videos and 1.5 when you need references, voices, pinned frames or native 1080p. Lite's draw is 480p: $0.02 a second against $0.08. At 1080p the saving shrinks, since Lite's upscaled 1080p lists at $0.14, the same as 1.5's 720p, against $0.25 for native.

Who it is for, and who should wait

Good fit:

• Clips where speech and sound should arrive with the picture, such as a talking shot, with no voiceover step.
• Product, character and location shots built from reference images.
• Animating a still at up to 1080p.

Test first, or wait, if:

• You need more than 15 seconds in one generation. The original row on Versely runs to 30.
• You need voices from your own audio, pinned frames or keyframes. Versely's catalog lists none of the three.
• You need reference shots above 720p.
• You edit or extend existing footage. xAI lists both on the original only.
• Price per second decides it. The original row costs about half the credits at 720p.

How to use it on Versely

As listed on 16 October 2026, Versely runs three 1.5 rows on one model page: Text to Video, Image to Video and Reference to Video. Each starts at 7 credits (one second at 480p) and bills by the second at 7, 12 and 20 credits for 480p, 720p and 1080p. Whole clips:

Clip	480p	720p	1080p
5 seconds	32	56	100
10 seconds	64	112	200
15 seconds	96	168	300

Those per-second rates are rounded up and a clip is priced once, so 10 seconds at 480p is 64 credits, not 70. Reference to Video stops at 720p, so its ceiling is 168 credits; the others reach 300. Text and Reference offer seven aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3), Image to Video follows your still, and Reference takes up to seven images, named <IMAGE_0>, <IMAGE_1> and so on in the prompt. There is no Lite row.

1. Open the AI video generator on the web or iPhone and pick a 1.5 row.
2. Set length and resolution first; 1080p is nearly three times the credits of 480p for each second.
3. Write the shot, the camera move and the sound. xAI's docs say to describe "the shot, including camera movement, lighting, and sound". More in the Grok Imagine prompting guide.
4. Test at 480p, then render the keeper at 720p or 1080p.
5. With references, xAI advises one role per image, a fixed place in the scene, a small cast and still props.

1.5 or the original? The original Grok Imagine Video row starts at 24 credits and runs 6, 10, 15, 20 or 30 seconds at up to 720p, and its Image to Video row prices the same. Ten seconds is 40 credits at 480p and 56 at 720p, against 64 and 112 on 1.5. Choose 1.5 for 1080p, lip-synced speech, references or clips under six seconds. Choose the original for longer takes or the fewest credits per second, and Grok Imagine Extend, from 13 credits, to continue a take.

FAQ

When was Grok Imagine Video 1.5 released?

xAI made it generally available in its API in June 2026; WinBuzzer reported it on 18 June. A 31 July 2026 update added text-to-video, native 1080p and reference-to-video.

How much does Grok Imagine Video 1.5 cost?

On xAI's API, as listed on 16 October 2026, 1.5 is $0.08 per second at 480p, $0.14 at 720p and $0.25 at 1080p; Lite is $0.02, $0.03 and $0.14. On Versely, 1.5 starts at 7 credits: 7, 12 and 20 per second at 480p, 720p and 1080p, up to 300 for 15 seconds at 1080p.

What is Grok Imagine Video 1.5 Lite?

xAI's cheaper text-to-video and image-to-video model, grok-imagine-video-1.5-lite: 1 to 15 seconds with lip-synced speech at 480p, 720p or 1080p upscaled from 720p, from $0.02 per second. It has no reference images, voices or pinned frames, and it is not in Versely's catalog.

How many reference images does Grok Imagine Video 1.5 take?

xAI's docs list up to 14 per request (7 on the original model) plus up to three preset voices, at up to 720p and 15 seconds. Versely's Reference to Video row takes up to seven images at 480p or 720p.

Can I use Grok Imagine Video 1.5 on Versely?

Yes. As of 16 October 2026 Versely lists Text to Video, Image to Video and Reference to Video rows, each from 7 credits, with native audio. Lite and voice references are not offered.