Skip to main content

Video Generation Cost Calculator

Sora per-second pricing by resolution — price a clip, a campaign, or a monthly pipeline.

100% client-side⛁ prices verified 2026-06-11⌁ zero network calls
10s
100
$1.00

per 10s clip on sora-2 · 720p$100.00/month at 100 clips/month.

Monthly cost across variants — same clip length & volume

sora-2 · 720p$100.00
sora-2-pro · 720p$300.00
sora-2-pro · 1024p$500.00
sora-2-pro · 1080p$700.00

Rates verified 2026-06-11. Effective rate: $0.10/second.

18
models priced, 4 vendors
2026-06-11
prices verified against vendor pages
90d
price staleness tripwire in CI
0
network requests per keystroke

How it works

This calculator prices Sora video generation the way OpenAI bills it: per second of output, with the rate set by model and resolution. Pick a variant, set your clip length and monthly clip count, and you get cost per clip and projected monthly spend — all computed locally in your browser with no signup.

The rate table is small but the spread is large. sora-2 at 720p runs $0.10 per second; sora-2-pro runs $0.30 at 720p, $0.50 at 1024p and $0.70 at 1080p. That means a 10-second clip costs anywhere from $1.00 to $7.00 depending on what you select — a 7× range for the same duration. Clip length compounds this linearly, so a team that habitually generates 20-second clips when 8 seconds would do is paying 2.5× more than it needs to on every single render.

The batch toggle applies OpenAI's Batch API discount, which halves every rate ($0.05, $0.15, $0.25 and $0.35 per second) in exchange for asynchronous turnaround of up to 24 hours. Most production video pipelines — catalog assets, ad variants, localized cuts — are not latency-sensitive, which makes batch the default choice rather than the optimization. The comparison bars show all four variants at your volume so you can see exactly what a resolution downgrade or a batch switch is worth per month.

One honest caveat: this prices accepted output. Real-world Sora usage involves re-rolls — generating multiple candidates and keeping one — and rejected generations still consume attempts. If your acceptance ratio is one keep per three generations, multiply the monthly figure by three. The calculator shows clean per-unit math; your pipeline determines the multiplier.

All rates were verified on 2026-06-11 against the OpenAI pricing page, and the verification date ships inside the tool's source as a documented constant. Video pricing is newer and less stable than text-token pricing, so check the date before putting these numbers in a budget. For measured spend across every model your team actually calls — video included — FORG tracks it from real session traffic. The share link preserves your exact scenario.

Frequently asked questions

How does Sora pricing work?

OpenAI bills Sora per second of generated video, with the rate set by model and resolution: sora-2 costs $0.10/sec at 720p, while sora-2-pro costs $0.30/sec at 720p, $0.50/sec at 1024p and $0.70/sec at 1080p. A 10-second 1080p pro clip is therefore $7.00 before any discount — clip length is the single biggest cost lever.

Is there a batch discount for video generation?

Yes. The Batch API halves every Sora rate — $0.05, $0.15, $0.25 and $0.35 per second respectively — in exchange for asynchronous processing with turnaround up to 24 hours. For marketing pipelines, A/B variant generation or any workload that renders overnight, batch pricing cuts the monthly bill in half with no quality difference.

Does resolution really triple the cost?

On sora-2-pro it more than doubles: 720p is $0.30/sec and 1080p is $0.70/sec, a 2.3× difference for the same clip length. If your clips are destined for social feeds where most viewers watch on phones, generating at 720p and upscaling in post is often indistinguishable — the comparison bars make the spread visible at your exact volume.

What is not included in this estimate?

Failed generations, content-policy rejections that still consume attempts, and regeneration loops where you re-roll a clip several times before accepting one. Teams routinely generate 2-4 candidates per published clip, so multiply the per-clip figure by your real acceptance ratio. Rates were verified 2026-06-11 against the OpenAI pricing page.

FORG tracks this automatically across every agent session — live cost attribution, budgets, and alerts.

Start tracking with FORG