Grok Imagine beside OpenAI Sora: delivery specs

Grok Imagine and OpenAI Sora both print figures on 3 spec lines, resolution, frame shape and length; their readings differ on frame shape and length. From 2 source pages, read 2026-09-25; no maker pays to be listed.

Fig. 1 Filled: the vendor prints a figure. Empty: no figure, or not read.

Grok Imagine and OpenAI Sora, reading by spec line, checked 2026-09-25.
Spec line	Grok Imagine	OpenAI Sora
Maximum resolution	1080p	1080p
Frame rate	No figure	No figure
HDR and bit depth	No figure	No figure
Aspect ratios	Three or more	Landscape and portrait
Longest single output	15 s or less	16 to 30 s
Watermark on output	No figure	No figure

Inclusion rule. The six spec lines every deliverable page carries. Order. Sheet order of the spec lines.

1Where the two sheets part

Grok prints seven ratios and 1 to 15 seconds; Sora prints landscape and portrait sizes and 16 or 20 seconds. Both reach 1080p, Grok on its 1.5 model for text and image to video.

Maximum resolution

Grok Imagine: 1080p, 720p or 480p (the default); 1080p on grok-imagine-video-1.5 for text and image to video. OpenAI Sora: 1080p exports, 1920x1080 or 1080x1920, on sora-2-pro; 1280x720 in the examples.

Aspect ratios

Grok Imagine: Seven shapes, from the default 16:9 through 3:2, 4:3 and square 1:1 to 3:4, 2:3 and 9:16 portrait. OpenAI Sora: Landscape and portrait sizes: 1920x1080 or 1080x1920 on sora-2-pro.

Longest single output

Grok Imagine: 1 to 15 seconds per video; edited videos keep their original length, capped at 8.7 seconds. OpenAI Sora: Up to 20 seconds; both models support 16- and 20-second generations.

2Figures on file

Resolution table
1080p, 720p or 480p (the default); 1080p on grok-imagine-video-1.5 for text and image to videoxAI docs, video generation / recorded 2026-09-25
Models
1080p exports, 1920x1080 or 1080x1920, on sora-2-pro; 1280x720 in the examplessora-2-proOpenAI API docs, video generation with Sora / recorded 2026-09-25
Aspect ratio table
Seven shapes, from the default 16:9 through 3:2, 4:3 and square 1:1 to 3:4, 2:3 and 9:16 portraitxAI docs, video generation / recorded 2026-09-25
Models
Landscape and portrait sizes: 1920x1080 or 1080x1920 on sora-2-prosizesOpenAI API docs, video generation with Sora / recorded 2026-09-25
Duration parameter
1 to 15 seconds per video; edited videos keep their original length, capped at 8.7 secondsxAI docs, video generation / recorded 2026-09-25
Choose size and duration
Up to 20 seconds; both models support 16- and 20-second generationsOpenAI API docs, video generation with Sora / recorded 2026-09-25

3One more delivery sheet beside Grok Imagine and OpenAI Sora

SceneMixer's printed figure on the checks where Grok Imagine and OpenAI Sora differ, read 2026-09-25.
Check	SceneMixer	Source
Aspect ratios	16:9 landscape by default, or 9:16 portrait, chosen once per project	SceneMixer, what comes out (video specs guide)
Longest single output	Segments of 15 or 30 seconds, chosen per episode; the tier table's longest clip is 30 seconds	SceneMixer, what comes out (video specs guide)

Inclusion rule. Checks on which the two tools differ and SceneMixer prints a figure; checks it prints nothing on are not shown. Order. The order checks take on every tool page.