Family holiday greeting: a 10-second Omni Flash card with music

Make a 10-second holiday greeting with Gemini Omni Flash 1.1 at 720p, a Lyria 3.5 music bed and a caption cue for the message. Chain and cost, read 2026-10-10.

4 min readSume
All posts

A 10-second holiday greeting costs about $1.78 on Sume: Omni Flash 1.1 at 720p ($1.25), one Lyria 3.5 music track ($0.125), one Timeline render ($0.10) and one caption job ($0.20) for the message, plus a $0.10 still if you want a reference scene. Ten seconds is the top of Omni's 3 to 10 second range.

A greeting is the one case where the video should be short, warm and have no spoken claim to get wrong. The model draws a scene (snow on a porch, a lit window), the music router makes a bed, and a cue prints the words. Prices come from the Sume pricing code (read 2026-10-10).

What is the endpoint chain?

The scene and the music have no dependency, so start both together.

  • POST /v1/images with a porch-at-dusk still, optional, to use as a reference image for Omni.
  • POST /v1/videos with gemini-omni-flash-1.1, duration: 10, resolution: 720p, aspect_ratio: 9:16.
  • POST /v1/music-router/generate with lyria-3.5, a gentle acoustic prompt and no vocals, or pick a free track from GET /v1/bgm/catalog at no cost.
  • POST /v1/timeline-1.0/render with the clip as slot one and soundtrack as the music; use audio.mode: "silence" with duration_seconds: 10.
  • POST /v1/video-captions with cues such as 'Happy holidays from the Reyes house'.

What does it cost?

Omni Flash is billed by output second and resolution. Compare the tiers for ten seconds.

Ten-second Omni Flash 1.1 by resolution (read 2026-10-10)
ResolutionPer second10 seconds
360p$0.0375$0.375
720p$0.125$1.25
1080p$0.1875$1.875
4K$0.375$3.75

What are the limits?

Omni Flash takes 16:9 or 9:16 and up to 10 reference images. The docs list image and video references and no audio input, so the music bed comes from Music Router or the free BGM catalog. The catalog is open without a key at /v1/bgm/catalog, /categories and /pick.

The full bill at 720p is: clip $1.25 + music $0.125 + render $0.10 + captions $0.20 = $1.675; with the free BGM track it is $1.55. Use Audio and music for the track options and Timeline 1.0 for soundtrack and fade_out_seconds (up to 10 seconds).

A greeting sent to twenty relatives is twenty copies of one render, not twenty jobs. Download once and share the file.

Is there a cheaper draft path?

Draft at 360p for $0.375, check the framing, then spend $1.25 for 720p. That is the same proof loop agencies use; see the 480p proof-then-final loop and the free background music catalog for the zero-cost bed.

Before you send it

Watch with the sound on and off. The caption cue carries the message with sound off, and the music should not fight with a spoken line if you add one later.

If the first draft shows an odd object in the scene, regenerate only the clip. The music and the caption are separate files and do not change.

  • Check the clip duration is between 3 and 10 seconds.
  • Set soundtrack.fade_out_seconds so the music ends softly; the cap is 10 seconds.
  • Save the final MP4 URL; it is the durable copy.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume