12-second Reel with an AI presenter: cost by quality tier

A 12-second talking-avatar Reel costs $2.208 on Sume's Standard tier, $2.94 on Plus and $6.60 on Max at the listed per-second rates, before any product-image add-on. That is a per-second price times the billable seconds, so the same arithmetic prices a 4-second hook or a 15-second punchline.

Instagram's Reels guide (read 2026-10-03) says the 3 to 15 second range is great for quick transitions, trends and punchy one-liners. That is the length where a presenter job is cheapest and where one strong line carries the whole Reel.

Rates and the 12-second price

The rates come from Sume's catalog rate card for sume/avatar-1.0/talking-video (rate card version sume_avatar_video_fixed_tier_rates_2026_07_02); the variants differ only when a product_image is sent.

What fixes the duration

Sume estimates the video duration from your script and accepts a job when that estimate lands between 4 and 60 seconds inclusive. Shorter scripts than 4 seconds are billed as 4 (the catalog reserves four billable seconds as the shortest chunk-planned clip), and longer ones must be split into several jobs. Output is 720p and the default aspect ratio is 9:16, which is what a Reel wants. Quality defaults to plus when you omit it.

Preview first, then pay for the render

For a line you have not tested, use the preview flow in Avatar video previews so you approve the still and the wording before the final render, and store caption intent on the preview. Inline captions are an option on the talking-video request: captions: { enabled: true, style: "slam", language: "auto" } burns the spoken script into the final MP4, they do not create a separate billed caption job, and an estimated duration above 60 seconds is rejected for inline captions. Read the quote on the request rather than assuming the total.

Choosing the tier for a 12-second Reel

Standard is the fastest path and the cheapest, so it suits a test of three hook lines; Plus is the balanced default; Max is the highest quality and slower, so keep it for the winner. Three 12-second hooks on Standard cost $6.624 in total at no-product rates, about the same as one 12-second Max render ($6.60). Sume does not change the presenter's performance after the fact: a new line or change is a new preview or render, and the earlier job is still billed.

What Sume does not do

It does not add trending audio or Instagram stickers, and it does not post the Reel; you download the MP4 from the job result and publish it yourself. Confirm live prices with GET /v1/catalog before a batch, because the rate card version above is dated and can be revised.