AI Image & Video Generation Cost Calculator
See what generating images or video actually costs, across every major provider — no subscriptions, no guessing.
Monthly cost by model
| Model | Provider | Price / image | Per month |
|---|
How this calculator works
Image generation is almost always billed per image, sometimes with separate rates for different quality or resolution tiers. Video generation is billed per second of output, with price scaling steeply by resolution — a model's cheapest and most expensive tiers are often 8-12× apart. Enter your monthly volume and this table multiplies it out and sorts cheapest first.
What actually moves the price
- Resolution/quality tier is the biggest lever, by far. Google's Veo 3.1 alone spans $0.05/second (Lite, 720p) to $0.60/second (Standard, 4K) — a 12× range on the same model family. The numbers shown here use each provider's standard/default tier; check the source link if you need an exact premium-tier quote.
- Video bills in whole seconds. Most providers round partial seconds up, so a 4.2-second clip is billed as 5 seconds.
- Batch/async discounts exist but aren't modeled here. OpenAI's Sora, for example, offers batch pricing at roughly half the standard rate for non-realtime workloads — the same pattern as the batch discounts on the LLM cost calculator, just not built into this table yet.
FAQ
Why isn't this data auto-synced like the LLM calculator?
The LLM pricing table syncs automatically from a public API (OpenRouter) that aggregates text-model pricing across providers. No equivalent public aggregator exists for image or video generation pricing — we checked before building this page. So this dataset is hand-verified against each provider's official pricing page instead, with the verification date and a source link on every row. It's the same approach the LLM calculator itself used before an auto-sync source existed for it.
Why do prices vary so much between resolution tiers?
Higher resolution and longer context (for video, more frames to generate consistently) cost providers more compute, and that gets passed through directly — video especially, since a 4K frame has roughly 4× the pixels of 1080p. If your use case can tolerate a lower tier, checking the tier breakdown on the provider's own pricing page can cut costs substantially.
Does this include subscription/credit-bundle plans?
No — this table only includes pay-as-you-go API rates. Several providers (Runway is a notable example) primarily sell consumer subscription plans with bundled credits rather than a published flat per-second API rate, which makes them hard to compare apples-to-apples here; we left those out rather than publish a number we couldn't directly verify.
How accurate are these numbers?
Each price was read directly from the provider's own pricing page on the date shown above the table — not estimated or recalled from memory. Generative AI pricing moves at least as often as LLM pricing does, so treat this as a snapshot and use the source link on any row to confirm the current rate before budgeting a large production run.