Monday, 9 November 2026 | Updating Daily AI insight, written for builders

AI Video Generator Cost Index (November 2026)

  • Cheapest API per second: Google’s Gemini Omni 1.2 Flash at $0.023 per second of 360p video, then MiniMax-H3-Lite at $0.025 (480P) on MiniMax’s own API.
  • Cheapest 5 seconds of 1080p through an API: Vidu Q3 Turbo at $0.35 ($0.20 off-peak), then Veo 3.1 Lite on Google Cloud at $0.40 with audio or $0.25 without.
  • Cheapest on a subscription: Google AI Ultra through Flow, $0.12 to $0.15 per 5-second clip, but upscaled from 720p. For video made at 1080p, Adobe Firefly’s paid plans run Veo 3.1 Lite for about $0.25.
  • Same model, different price: 5 seconds of Kling 4.0 at 1080p costs from $0.79 on Luma Ultra to $2.83 on Runway Standard.
  • 54 rows of official US prices, checked 9 November 2026. Free to reuse under CC BY 4.0 as CSV or JSON.

How much does an AI video generator cost?

In November 2026, AI video generator cost runs from $0.023 per second of video through an API to about $1.16 per second for Seedance 2.5 at 1080p on fal, and paid plans start at $4.99 a month. For 5 seconds of 1080p, Google AI Ultra is the cheapest subscription route ($0.12 to $0.15, upscaled from 720p) and Vidu Q3 Turbo the cheapest API, at $0.35.

This index converts the published prices of 14 AI video tools, plus OpenAI’s retired Sora 2, into one unit: the cost of 5 seconds of 1080p video. Every figure comes from the vendor’s own US pages, or from fal and BytePlus for hosted models, checked on 9 November 2026. Rankings reflect published prices, not output quality.

The index: what 5 seconds of 1080p costs

Sorted by the cost of 5 seconds of 1080p video. “Not published” means the vendor gives no 1080p rate or credits-per-clip figure.

Tool Plan or API Price Model priced Lowest API rate per second 5 s at 1080p
Google (Flow) Google AI Ultra, $199.99 tier $199.99/mo Omni Flash 1.2 n/a $0.12 (720p upscaled)
Google (Flow) Google AI Ultra $99.99/mo Omni Flash 1.2 n/a $0.15 (720p upscaled)
Adobe Firefly Standard $9.99/mo Veo 3.1 Lite n/a $0.25
Adobe Firefly Pro $19.99/mo Veo 3.1 Lite n/a $0.25
Adobe Firefly Pro Plus $49.99/mo Veo 3.1 Lite n/a $0.25
Google (Flow) Google AI Pro $19.99/mo Omni Flash 1.2 n/a $0.30 (720p upscaled)
Vidu Q3 Turbo API Per second Vidu Q3 Turbo $0.04 (540p) $0.35 ($0.20 off-peak)
Google (Flow) Google AI Plus $4.99/mo Omni Flash 1.2 n/a $0.37 (720p upscaled)
Google Veo 3.1 Lite, Google Cloud (preview) Per second Veo 3.1 Lite $0.03 (720p, no audio) $0.40 with audio, $0.25 without
Luma Ultra $300/mo Kling 3.0 (with audio) n/a $0.49
Google Gemini Omni 1.2 Flash API Per token Omni 1.2 Flash $0.023 (360p) $0.52 (upsampled)
Luma Pro $90/mo Kling 3.0 (with audio) n/a $0.55
Google Veo 3.1 Fast, Google Cloud Per second Veo 3.1 Fast $0.08 (720p, no audio) $0.60 with audio
Vidu Q4 Preview API (image or reference to video) Per second Vidu Q4 Preview $0.045 (540p) $0.60
Luma Plus $30/mo Kling 3.0 (with audio) n/a $0.74
Wan Wan 2.5 (fal) Per second Wan 2.5 $0.05 (480p) $0.75
Kling Kling 4.0 API (Kling) Prepaid units Kling 4.0 $0.157 (720P) $1.18
Kling Kling 4.0 (fal) Per second Kling 4.0 $0.157 (720p) $1.18
Kling Kling 4.0 (Runway API) Per second Kling 4.0 $0.157 (720p) $1.18 ($0.94 to 30 Nov)
Runway Max $95/mo Kling 4.0 n/a $1.18
Luma Ray3.2 API Per video Ray3.2 $0.03 (540p, $0.15 per 5 s) $1.20
Pika Fancy $95/mo Kling 4.0 n/a $1.31
Pika Creator $35/mo Kling 4.0 n/a $1.59
Runway Pro $35/mo Kling 4.0 n/a $1.84
Pika Starter $10/mo Kling 4.0 n/a $1.97
Google Veo 3.1, Google Cloud Per second Veo 3.1 $0.20 (720p/1080p, no audio) $2.00 with audio
Runway Standard $15/mo Kling 4.0 n/a $2.83
Seedance Seedance 2.5 API (BytePlus) Per token Seedance 2.5 $0.103 (480p) $2.84
Seedance Seedance 2.5 (fal) Per token Seedance 2.5 $0.2205 (480p) $5.82
MiniMax MiniMax-H3-Lite API Per second MiniMax-H3-Lite $0.025 (480P) Not published (no 1080P)
MiniMax MiniMax-H3 API Per second MiniMax-H3 $0.08 (768P) Not published (no 1080P)
Kling Kling 4.0 Flash (fal) Per second Kling 4.0 Flash $0.094 (720p) 720p only
Runway Gen-4.5 API Per second Gen-4.5 $0.12 (720p) 720p only
Kling Standard $8.80/mo n/a n/a Not published
Kling Pro $32.56/mo n/a n/a Not published
Adobe Firefly Premium $199.99/mo Firefly Video Model n/a Unlimited own-model use
HeyGen Creator $29/mo n/a n/a Not published
Synthesia Starter $19/mo n/a n/a Not published
Magic Hour Creator $19/mo n/a n/a Max 1024px
Vidu Standard $10/mo n/a n/a Not published
PixVerse Standard $8/mo, billed yearly n/a n/a Max 720P
OpenAI Sora 2 Pro API (shut down 24 Sep 2026) Per second sora-2-pro $0.30 (720p) $3.50 (historical)

Cheapest options by type

Cheapest API per second

Google’s Gemini Omni 1.2 Flash has the lowest standard rate in the index: $0.023 per second of 360p video. Google bills it as video output tokens at $12 per million, about $0.07 per second at 720p, $0.104 at 1080p and $0.209 at 4K (the last two upsampled), on the paid tier only.

MiniMax’s own API is close behind. MiniMax-H3-Lite costs $0.025 per second at 480P and $0.04 at 768P, and the flagship MiniMax-H3 costs $0.08 at 768P and $0.13 at 2K. Vidu’s off-peak rate for Q3 Turbo, $0.02 per second at 540p, is lower still, but Vidu calls all its API prices estimates and says off-peak eligibility varies by model.

Cheapest 1080p through an API

At 1080p, Vidu Q3 Turbo lists $0.07 per second: $0.35 for 5 seconds, or $0.20 off-peak. Veo 3.1 Lite on Google Cloud lists $0.08 per second with audio and $0.05 without, so 5 seconds costs $0.40 or $0.25. Google still labels Lite a preview there, while Veo 3.1 and Veo 3.1 Fast are generally available.

Cheapest clip on a subscription

Google’s AI plans give the lowest per-clip figure of any subscription. Flow charges 3 credits per second for Omni Flash 1.2 at 720p, and upscaling a 1- to 10-second clip to 1080p costs 0 credits for Plus, Pro and Ultra subscribers. Five seconds uses 15 credits: about $0.37 on Google AI Plus ($4.99 for 200 monthly credits), $0.30 on AI Pro ($19.99 for 1,000), $0.15 on AI Ultra ($99.99 for 10,000) and $0.12 on the $199.99 Ultra tier (25,000). Every Flow user also gets 50 credits a day, which the index leaves out. The video is made at 720p and upscaled.

For video made at 1080p, Adobe Firefly’s paid plans are the cheapest route, because they include Google’s Veo 3.1 Lite as a partner model at 10 credits per second at 1080p. Five seconds uses 50 credits, about $0.25 on Standard ($9.99 for 2,000 credits), Pro or Pro Plus. Adobe’s own Firefly Video Model uses 100 credits per second, or $2.50 for the same 5 seconds on Standard. Adobe calls its own models’ output safe for commercial use but leaves partner-model output to you. Our Adobe Firefly guide covers the plans.

Luma is next for native 1080p: Kling 3.0 with audio, its cheapest 1080p model at 49 credits per second, costs about $0.49 per 5 seconds on Ultra and $0.74 on Plus.

One model, many prices: Kling 4.0

Kling 4.0 is sold almost everywhere, which shows what the storefront adds. Here is 5 seconds of Kling 4.0 at 1080p at each seller’s list price:

Where you buy it How it is priced 5 s of Kling 4.0 at 1080p
Luma Ultra ($300/mo) 79 credits per second $0.79
Luma Pro ($90/mo) 79 credits per second $0.89
Kling API (Kling’s developer platform) $0.235 per second (1.5 prepaid units) $1.18 ($1.15 with the $500 package)
fal $0.235 per second $1.18
Runway API 23.5 credits per second at $0.01 $1.18 ($0.94 until 30 November)
Runway Max ($95/mo) 118 credits per 5 seconds $1.18
Luma Plus ($30/mo) 79 credits per second $1.19
Adobe Firefly Standard, Pro or Pro Plus 50 credits per second $1.25
Pika Fancy ($95/mo) 118 credits per 5 seconds $1.31
Pika Creator ($35/mo) 118 credits per 5 seconds $1.59
Runway Pro ($35/mo) 118 credits per 5 seconds $1.84
Pika Starter ($10/mo) 118 credits per 5 seconds $1.97
Runway Standard ($15/mo) 118 credits per 5 seconds $2.83 ($2.14 until 30 November)
Kling’s own plans Credits per clip not published Not published

The dearest seller charges 3.6 times what the cheapest does for the same clip. Entry plans are the dearest per clip: Runway Standard and Pika Starter charge more for 5 seconds of Kling 4.0 than fal or Kling’s own API, because a $10 or $15 plan buys few credits. Those plans add other models and creative tools, which an API does not. For Kling’s own plans, see our Kling AI guide.

Where vendors do not publish a per-clip cost

Several popular tools sell credits or minutes without saying what 5 seconds of 1080p uses, so the index leaves those values empty.

  • Kling’s own plans list 660 credits a month on Standard ($6.99 for the first month, then $8.80) and 3,000 on Pro ($25.99, then $32.56), but not the credits a Kling 4.0 clip uses.
  • HeyGen says the credits a video uses vary with the model, duration and complexity. Its help center gives only a “starting at” rate for generated video, which varies by model, resolution and length. Our HeyGen pricing guide explains the credits.
  • Synthesia sells avatar-video minutes: up to about 12 a month on Starter, on top of 10 free minutes. The figure is approximate, so it is not converted.
  • Magic Hour exports up to 1024px on Creator, short of 1080p.
  • PixVerse and Vidu: PixVerse Standard gives HD up to 720P, and Vidu’s plan cards publish only a 540P rate (2 credits per second for its Q4 model).
  • MiniMax publishes no 1080P rate for H3: its API goes from 768P to 2K. Hailuo’s consumer plan page would not load in readable form, so the Hailuo app is left out this month.

Retired and retiring models

OpenAI shut down the Videos API and the sora-2 and sora-2-pro models on 24 September 2026 and named no replacement. Their last list prices stay as historical rows: sora-2 at $0.10 per second at 720p, and sora-2-pro at $0.30, $0.50 and $0.70 per second at 720×1280, 1024×1792 and 1080×1920. The Sora 2 API shutdown explainer has the background.

Google shut down the Veo 3.1 preview models in the Gemini API on 22 October 2026 and names gemini-omni-1.2-flash as their replacement. Veo 3.1 and Veo 3.1 Fast remain generally available on Google Cloud, with retirement listed as “May 4, 2027 or later”. More changes are scheduled:

  • 17 November 2026: fal deprecates its Veo 3.1 endpoints.
  • 30 November 2026: Runway’s Kling 4.0 discounts end (25% off on plans, 20% off on its API).
  • 3 December 2026: requests to Gemini Omni 1.1 Flash start routing to 1.2 Flash, billed at 1.2 rates.
  • 5 December 2026: Flow retires Veo 3.1 Fast, Veo 3.1 Quality and Omni Flash 1.1.

How we calculated cost per clip

All figures use official US list prices in US dollars, before tax.

  • Subscriptions: the monthly price on monthly billing, divided by the plan’s monthly credits, times the credits the vendor says 5 seconds at 1080p uses, with the cheapest 1080p model the vendor publishes a rate for (the rate with audio, where audio is priced separately). First-month discounts, promotions and free daily credits are left out.
  • Google’s plans: Omni Flash 1.2 at 720p, upscaled to 1080p, which Flow prices at 0 extra credits for clips up to 10 seconds. Flow’s Veo 3.1 Lite costs fewer credits per clip but is listed at 720p only, with no published upscale rate.
  • APIs: five times the 1080p per-second rate. Where audio is priced separately, the index uses the rate with audio and gives the silent price in the method field. Where a vendor publishes its own 5-second 1080p price, as Luma (per video) and BytePlus (a worked example) do, the index uses it.
  • Fixed clip lengths: Veo makes 4-, 6- or 8-second clips, so its 5-second figure is a per-second unit, not a clip you can order.
  • Google Cloud: Google lists Veo prices per “count”. Its own SKU catalog bills the same Veo SKUs at the same prices per second, so the index reads a count as one second. Google also lists pay-as-you-go as not supported for Veo 3.1 on Google Cloud; access runs on fixed quota or Provisioned Throughput.
  • Empty values: no published 1080p rate or credits-per-clip figure means null, with the reason in the method field.
  • One take only: every figure is the cost of one successful generation. If a usable shot takes four attempts, multiply by four.

Download the data

The full index is free to reuse under the Creative Commons Attribution 4.0 (CC BY 4.0) license: you may share and adapt it for any purpose, even commercially, as long as you credit Convly AI, link to the license and indicate any changes.

Both files hold the same 54 rows and 13 fields, free plans included, with each cost’s formula and source link.

Cite this index

Convly AI, AI Video Generator Cost Index, November 2026, https://convly.ai/ai-video-cost-index/

Choosing a tool rather than a price? See the best AI video generators, the best free AI video generators and the AI video generation pricing guide, plus Google’s AI plans explained and how to use Runway. Single models: MiniMax H3, Seedance 2.5, Veo 3.1 and Wan 2.5.

FAQ

What is the cheapest AI video generator API per second?

Google’s Gemini Omni 1.2 Flash, at $0.023 per second of 360p video on the paid tier of the Gemini API. MiniMax-H3-Lite is next at $0.025 per second at 480P.

How much does a 5-second 1080p AI video cost?

Through an API, from $0.35 (Vidu Q3 Turbo) to $5.82 (Seedance 2.5 on fal) for one successful generation. On a subscription, from about $0.25 (Veo 3.1 Lite inside Adobe Firefly) to $2.83 (Kling 4.0 on Runway Standard), or $0.12 on Google AI Ultra if 720p upscaled is acceptable.

Which AI video subscription gives the cheapest clips?

Google AI Ultra through Flow, at $0.12 to $0.15 per 5-second clip, upscaled from 720p. For video made at 1080p, Adobe Firefly’s Standard, Pro and Pro Plus plans run Veo 3.1 Lite for about $0.25 per 5 seconds.

Why does the same model cost different amounts?

Each platform sets its own credit price. Five seconds of Kling 4.0 at 1080p costs from $0.79 on Luma Ultra to $2.83 on Runway Standard, and small plans cost more per clip because they buy fewer credits.

Is Sora 2 still in the index?

Only as history. OpenAI shut down the Videos API and the sora-2 and sora-2-pro models on 24 September 2026 with no replacement; their last list prices, from $0.10 per second, are kept as historical rows.

Can I reuse this data?

Yes, under CC BY 4.0: you can share and adapt the CSV and JSON files, even commercially, if you credit Convly AI, link to the license and indicate any changes. The “Cite this index” box has a ready-made citation.

Written by Mustafa Ihsan

Mustafa Ihsan is the founder and editor of Convly.ai. He built and maintains the site's live AI models database, its price-performance index, and its free calculators for VRAM requirements, API costs and self-hosting economics. He writes about model pricing, benchmark results and the hardware needed to run AI models locally, and consistently prefers measured numbers to vendor claims.

Scroll to Top