How AI video API pricing works
AI video APIs are usually billed in one of two ways: per generated second of video, or through a monthly allowance of renders. Avatarity uses monthly allowances for Veo 3.1 Lite and Nano Banana, with Omni 1.1 Flash as an optional pay-per-video add-on. This page explains what that means for a production pipeline.
Per-second billing and monthly allowances
With per-second billing, every generated second is charged, including takes that never make the final edit. Costs follow your volume directly and are easy to start with, but harder to plan when a pipeline makes many drafts.
With a monthly allowance, you pay a fixed price for a set number of renders. Costs are known in advance; the trade-off is that unused renders do not carry over to the next month.
Avatarity plans
- Starter $30/mo
- 2,000 Veo 3.1 Lite videos and 5,000 Nano Banana images per month, 4 renders at a time
- Pro $70/mo
- 8,000 Veo 3.1 Lite videos and 20,000 Nano Banana images per month, 4 renders at a time
- Advanced $250/mo
- 25,000 Veo 3.1 Lite videos and 50,000 Nano Banana images per month, 8 renders at a time
- Dedicated $400/mo
- One private dedicated render node, unmetered (fair use), 16 renders at a time
- A 720p Veo 3.1 Lite clip (about 8 seconds) counts as 1 video; a direct 1080p clip counts as 2
- Upscaling a finished 720p clip with
source_task_idcounts as 1 more video - Each Nano Banana image counts as 1 image
- The allowance resets every paid month and does not roll over
Worked examples
- A daily Shorts channel with 30 clips a day uses about 900 videos a month: within Starter (2,000 videos).
- Ten 10-minute videos a month at about 100 clips each (75 used plus 25 extra takes) use 1,000 videos: within Starter, with Pro leaving room for 1080p upscales.
- An agency producing for several channels at 6,000 clips a month plus thumbnails fits Pro (8,000 videos, 20,000 images).
Not sure which plan fits? Compare them on the pricing page or ask support.
Failed renders and retries
A render reserves one unit while it runs. If it ends as failed or cancelled, the unit goes back to your allowance automatically. Send an Idempotency-Key header with each request: a retry with the same key within 24 hours returns the original task instead of starting (and counting) a second render.
Questions and answers
What does a monthly plan include?
A fixed number of Veo 3.1 Lite videos and Nano Banana images per month and a number of renders that can run at the same time: Starter ($30) 2,000 videos and 5,000 images; Pro ($70) 8,000 videos and 20,000 images; Advanced ($250) 25,000 videos and 50,000 images; Dedicated ($400) one private node, unmetered (fair use). Prices in USD, net.
Is API access extra?
No. The same plan allowance works in the Studio and through the API with your Bearer key.
Do failed renders or retries cost anything?
Failed and cancelled renders are not counted. A retry that sends the same Idempotency-Key within 24 hours returns the original task instead of starting a new render.
Is there a pay-per-video option?
Omni 1.1 Flash is an optional add-on at $0.15 per successful generation, billed with the next monthly invoice up to a monthly spending cap you set. Failed generations are not billed.