Omni Flash native audio vs Wan 3.0 plus TTS and Timeline: cost

Gemini Omni Flash 1.1 on Sume makes video with native audio always on, in clips of 3 to 10 seconds. Wan 3.0 is billed per second by resolution, and a clip you want spoken needs voice from somewhere else. This post prices the two routes for a 10-second clip with a 400-character script, using Sume billed rates read 2026-10-03.

The comparison is about money and steps. It does not claim that one produces better speech or picture than the other.

What does each route cost?

Omni at 720p is $0.125 per second, so ten seconds is $1.25 and includes audio. The Wan route is a $1.25 clip at the same resolution, plus text-to-speech at the Sume rate of $0.0475 per 1,000 characters (47.5 micro-dollars per character, from the provider-pricing rate card), plus a Timeline 1.0 render at $0.10 per output minute, rounded up.

Is the Timeline step necessary?

Timeline 1.0 takes one audio spine and ordered video slots and returns an MP4, so it is the documented way to put a TTS voice-over under a Wan clip. Omni skips that step because its audio is part of the generation. The Timeline render is billed per whole output minute (Timeline 1.0), so a 10-second clip pays for a full minute.

If you assemble several clips in one Timeline, the per-minute charge is shared, which narrows the gap further. A one-minute piece of six 10-second slots pays the same $0.10 once.

How does it change at 1080p?

At 1080p, Omni is $0.1875 per second ($1.875 for 10 seconds), while Wan is $0.25 per second ($2.50). The Omni clip is about 28 percent cheaper than the full Wan route. At 720p the two clip prices are equal and the Wan route is dearer only because of its extra steps, about $0.12.

Omni clips top out at 10 seconds. For anything longer, you stitch or use Wan's 30-second window, which is where the second route earns its extra steps. See the single-take versus stitched comparison.