10 second talking head cost: Avatar 1.0, Fabric and H3 Max lip sync

A 10 second talking head costs between $0.625 and $5.50 in Sume, depending on the route. Avatar 1.0 starts from a script and costs $1.84 (standard), $2.45 (plus) or $5.50 (max) with no product image. Fabric and H3 Max lip sync start from audio you already have and a still image, and bill per audio second.

The ladder

Rates are from the Sume API pricing page, multiplied by 10 seconds.

Which input you have

If you have only words, Avatar 1.0 makes the voice and the face together. If you already recorded audio, Fabric or H3 Max animate a still to it. Audio for these routes must be Sume-hosted: under 10 MB, with Fabric allowing up to 300 seconds and H3 Max 5 to 14.8 seconds.

Text to speech adds $0.0475 per 1,000 characters if you generate the audio in Sume first.

Choose by need

Avatar 1.0 gives a consistent presenter across clips and optional captions. Lip sync routes give the most control over your own audio and the lowest per-second price at low resolution. H3 Max is limited to 5 to 14.8 seconds, so a 30 second clip needs another route.

Render a preview or a short test before running a batch, and compare faces, not only prices.