AI video ratios on Sume: Omni Flash is 16:9 or 9:16, Kling adds 1:1
Gemini Omni Flash 1.1 on Sume takes only 16:9 and 9:16; Kling 3 adds 1:1. Which other Sume video models list more ratios, and which list none.
On Sume, gemini-omni-flash-1.1 lists only two aspect ratios, 16:9 and 9:16. kling-3 adds a third, 1:1. Every other video row that takes an aspect ratio lists at least seven values, and three rows list none at all. If you need a 4:3 or 21:9 clip, pick a different model, because a value outside a row's list is not part of that row's contract.
Everything here is from the video router catalog in Sume's repo and its Video Router and Video generation docs, read 2026-10-03. The docs tell you to read capabilities from GET /v1/video-router/models instead of assuming one envelope, and that is still the safest check.
Ratios by model
The count includes the auto and adaptive values, which follow the source image or frame instead of fixing a shape.
| Id | Ratios listed | Count |
|---|---|---|
| gemini-omni-flash-1.1 | 16:9, 9:16 | 2 |
| kling-3 | 16:9, 9:16, 1:1 | 3 |
| wan-3.0 | auto, adaptive, 16:9, 4:3, 1:1, 3:4, 9:16 | 7 |
| minimax-h3 and minimax-h3-max | adaptive, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 | 7 |
| seedance-2.5, seedance-2, seedance-2-fast, seedance-2-mini | auto, adaptive, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 | 8 |
| grok-imagine-video-1.5, higgsfield-genjutsu, h3-max-recast | none (no aspect_ratio field) | 0 |
What the two-ratio limit means in practice
Omni Flash is a good fit for the two shapes most platforms want: landscape for YouTube and vertical for Reels, Shorts and TikTok. It is not a fit for a 1:1 feed post. Its other limits are 3 to 10 seconds, resolutions of 360p, 720p, 1080p and 4K, and native synced audio that is always on.
Kling 3 adds the square at 720p or 1080p for 4 to 15 seconds, with an optional generate_audio toggle. It still has no 4:3, 3:4 or 21:9.
Picking when you need a specific shape
On the rows that list adaptive or auto, the first frame can decide the shape, which avoids a mismatch between a still and the requested ratio.
- Square (1:1): Kling 3, Wan 3.0, MiniMax H3 and H3 Max, or any Seedance row.
- 4:3 or 3:4: Wan 3.0, the MiniMax H3 rows, or the Seedance rows.
- 21:9 ultrawide: the MiniMax H3 rows or the Seedance rows only.
- No ratio choice (the output follows your source): Grok Imagine Video 1.5, Genjutsu and H3 Max Recast.
What Sume does and does not do
Sume publishes each row's ratio list in the catalog. It does not make one model serve every shape. If a campaign needs both 16:9 and 1:1 from one row, choose a row that lists both, or finish in a Timeline where you control the output shape.
Sources
Related posts
More in Models
- Video-to-video clip length limits on Sume, by tool
Recast takes 5 to 30 s, Genjutsu 4 to 30 s, Omni edit 3 to 10 s, Avatar face swap about 4 to 15 s. A table and a Python picker choose by clip length.
- Vietnamese, Thai, Indonesian, Malay TTS API: vi, th, id, ms on Sume
Cartesia Sonic 3.6 lists vi, th, id and ms. How to send each as the language on Sume TTS, why the field is required, and a per-character cost for each script.
- What happens to the audio in each Sume video-to-video tool
Recast and Avatar Face Swap keep the source audio, Kling motion control keeps it by default, Omni always produces native audio, Genjutsu takes no audio field.
- Where to try MAI-Voice-2.1 before paying: a 10-minute listening test
Microsoft lists the MAI Playground, Copilot Audio Expressions and Foundry for MAI-Voice-2.1. Run a ten-minute listening test with a fixed script and cost it.
Written by Sume