Sound included: 8-second AI video clips at or under $1.00 on Sume

Six tiers on three Sume models produce an 8-second clip with built-in audio for $1.00 or less: Gemini Omni Flash 1.1 at 360p ($0.30) and 720p ($1.00), MiniMax H3 at 480p ($0.50) and 768p ($0.60), and MiniMax H3 Max at 480p ($0.50) and 768p ($0.80). On all three, audio is always on; Omni's API rejects generate_audio: false, and H3 and H3 Max have no toggle and always produce native stereo audio.

Prices for 8 seconds

Every figure is the Sume per-second rate multiplied by 8. Higher tiers are left out because they exceed the $1.00 line: Omni 1080p is 8 x $0.1875 = $1.50 and H3 Max 1080p is 8 x $0.20 = $1.60.

What differs besides price

Omni accepts 3 to 10 seconds in 16:9 or 9:16 only, while the two MiniMax rows accept 5 to 15 seconds and ratios from 21:9 to 9:16. The 768p tier on both MiniMax rows is native 768p, not 720p. H3 Max also offers 1080p, refined from native 768p.

Two Sume rows left out of the table are not audio-on-always. Seedance and Wan expose audio as a capability that generate_audio controls, and Grok Imagine Video 1.5 is silent. The pricing for Seedance and Wan in this post uses default options, and the docs do not describe a separate audio surcharge.

• Omni: 3 to 10 s, 16:9 or 9:16, 360p to 4K.
• H3 and H3 Max: 5 to 15 s, native stereo audio, 21:9 to 9:16.
• If you need silence, generate with sound and strip it, or use a silent row.

Picking one

For a clip of 10 seconds or less that may need 4K later, Omni at 720p is $1.00 for 8 seconds. For a clip that needs a wider ratio such as 21:9, or more than 10 seconds, H3 at 768p is $0.60 for 8 seconds, the cheapest tier of 768p or higher in the list above.

If you need to remove the sound

None of the three rows lets you turn audio off, so a silent clip needs one more step. Sume has an audio-detach job at $0.01 per job to pull the track out, and a Timeline can lay a different audio spine under the clip.

Plan for it at the budget stage: an Omni 720p clip at $1.00 plus a $0.01 detach is $1.01, which is over the $1.00 line. For a silent deliverable under a dollar, Grok Imagine Video 1.5 at $0.0125 per second (8 seconds is $0.10) is silent by design, but it needs a start image.