AI image & video generation API pricing
Images from $0.02. Video from $0.044 per second of output. Prepaid credits, no subscription, and the exact cost on every response.
Image generation
Priced per image, set by tier. On the fast and balanced tiers the price is flat — it never changes based on which model served it. Premium starts at $0.06 and can vary with the model; the exact charge is on every response.
| Tier | Quality | Price |
|---|---|---|
| corent-image-fast | Standard | $0.02 / image |
| corent-image-balanced | Standard | $0.05 / image |
| corent-image-premium | Premium | from $0.06 / image |
Video generation
Video bills per second of output, so a 5-second clip costs half of a 10-second one at the same rate. Rates are "from" prices — the per-second rate depends on which model the router conducts inside your tier, and the exact charge comes back on every response. No GPU-time math to reverse-engineer.
| Tier | Quality | Price |
|---|---|---|
| corent-video-fast | Standard | from $0.044 / second ≈ $0.27 per 6s clip |
| corent-video-premium | High | from $0.062 / second ≈ $0.31 per 5s clip |
| corent-video-cinematic | Premium | from $0.161 / second ≈ $0.81 per 5s clip |
Or just say how to optimize
Instead of naming a tier, pass a one-word preference with your request. This is how it resolves — and what that costs:
| Preference | Images | Video |
|---|---|---|
| fast | $0.02 | from $0.044/s |
| cheap | $0.02 | from $0.044/s |
| balanced | $0.05 | from $0.062/s |
| quality | from $0.06 | from $0.161/s |
Prices reflect our cost plus margin and can change over time. You always see the exact charge in the API response and your usage history.
How prepaid credits work
Top up
Add prepaid credit from $5 by card — it lands in your balance instantly. Crypto top-ups start at $10.
Generate
Each request checks your balance, runs, and debits the exact cost only after the result completes.
See the receipt
The exact charge comes back on every response, and every debit is a line item in your dashboard.
Failed generations are never charged. Corent retries across models automatically; if everything fails, the job is marked failed and your balance is untouched. You pay for pixels, not attempts.
Pricing questions, answered
How does Corent's AI image and video API pricing work?
Corent is pay-as-you-go with prepaid credits. You add credit to your account (from $5 by card), then each generation debits the exact cost of that result after it completes. There is no subscription and no monthly fee. Every API response includes the exact amount you were charged.
How much does an AI-generated image cost?
Images cost $0.02 on the fast tier, $0.05 on the balanced tier, and from $0.06 on the premium tier. You pick a tier (or a preference like fast, cheap, balanced, or quality) and Corent routes your prompt to the best available model in that tier.
How much does AI video generation cost per second?
Video is billed per second of output, and the per-second rate depends on which model the router selects inside your tier. Fast starts at $0.044/s, premium at $0.062/s, and cinematic at $0.161/s, so a 5-second clip starts around $0.22. The rate rises when the router reaches for a stronger model, and the exact charge for the model it picked is on every response. Failed generations are never billed.
Do I pay for failed generations?
No. Your balance is only debited after a generation completes successfully. If every underlying provider fails, the job is marked failed and you are charged nothing. Corent also retries across models automatically before a request is allowed to fail.
Is there a subscription, minimum spend, or platform fee?
No subscription and no platform fee. The only minimum is the smallest top-up: $5 by card ($10 for crypto payments). After that, you spend your balance in cents per generation, at the prices listed on this page.
How do I know exactly what I was charged?
Every API response includes the exact charge for that generation, and your dashboard shows a line item for every debit. You never have to reverse-engineer GPU-seconds or credit math — the receipt is part of the product.
Do I choose the model, or does Corent?
You choose a tier or a preference (fast, cheap, balanced, quality); Corent picks the best model behind it and reroutes if a provider fails. Fast and balanced image tiers are flat prices. On the 'from' tiers — premium images and all video — the rate depends on the model the router selects, and the exact charge is returned on every response and itemized in your dashboard.
Start with $5, keep the receipts
One API key, every tier. Read why routing beats picking a model, or grab a key and generate something.
Get started