How most AI video APIs charge
Unlike chat models (tokens) or image APIs (per image), video models often price by seconds of generated video, sometimes with separate rates for resolution or quality tiers. Your real bill is:
rate × seconds × clips × (1 + retry%)
Flat "one video = $X" thinking underestimates cost once you iterate. Three takes of an 8-second clip are three billable generations.
Example: Google Veo 3.1
This site specializes in Veo 3.1 Gemini API pricing. Snapshot as of Oct 9, 2026: Standard $0.40/s (720p–1080p), Fast $0.12/s at 1080p, Lite $0.08/s at 1080p. An 8-second 1080p clip is about $3.20 / $0.96 / $0.64 before retries.
Dive deeper: Veo 3.1 cost calculator (homepage), cost per second, Lite pricing, 4K cost, Fast vs Standard.
Free Veo 3.1 calculator
Use the tool below to estimate Veo 3.1 API spend by tier, resolution, length, and retry waste — then open the full homepage calculator experience for pricing tables and FAQ.
Estimate your Veo 3.1 bill
Pick tier, resolution, and how many takes you expect (including waste).
Estimated cost
$1.15
≈ 9.6 billable seconds @ $0.12/sec
- One clip
- $0.96
- Clips × waste
- 1 × 1.20
- Tier / res
- fast · 1080p
Rates reflect publicly listed Gemini API Veo 3.1 prices (with audio). Google may change pricing; always confirm on ai.google.dev.
Cite this (quick facts)
Plain numbers for articles, briefs, or AI assistants · Updated Oct 9, 2026
Veo 3.1 on the Gemini API is billed per output second (with audio). As of Oct 9, 2026: Standard is $0.40/s at 720p–1080p and $0.60/s at 4K; Fast is $0.10/s (720p), $0.12/s (1080p), $0.30/s (4K); Lite is $0.05/s (720p) and $0.08/s (1080p), no 4K. An 8-second 1080p clip costs about $3.20 (Standard), $0.96 (Fast), or $0.64 (Lite) before retries. There is no free Veo API tier. Source: https://www.veopricing.com