Veo 3 cost per second: Standard, Fast, and Lite in dollars

Kling credits live on Kling AI pricing explained. The Kling-vs-Veo compare lives on Kling vs Veo cost. This page does not redo that compare.

Veo 3.1 at a glance

How to read this table

• USD figures are Gemini API, paid tier, per second, “video with audio price (default)”, fetched 2026-09-18. Gemini does not list a silent Veo 3.1 rate. Vertex prints a second table (silent vs with-audio) in the next H2s. Model ids are what Google printed on that day, not nicknames.
• Gemini copy: Paid Tier, per second in USD. Free tier for these Veo 3.1 rows: Not available.
• Gemini note on the same block: “You will only be charged if your video is successfully generated.” An audio processing issue can block a take. A take that succeeds and is unusable still costs the full clip.
• Vertex prints the same with-audio dollars as $X / 1 count. This page treats 1 count as 1 billed second because the dollars match Gemini’s per-second table when checked. Vertex does not print the word “second” on that Veo grid. If the unit is no longer a second, stop and re-read.
• Vertex Standard and Fast cards on Veo 3.1 generate (Last-Modified 2026-09-16): Standard launch stage GA (release 2025-11-17). Lite launch stage Preview (release 2026-04-02). Fast’s own versions block is GA; a Preview notice sits on the same page above Lite. Confirm stage on the card you will call.
• Gemini generate docs still tag Veo 3.1 and Veo 3.1 Lite as Preview. Vertex uses -001 ids. Gemini uses -preview ids. Same family, two SKUs. Paste the id your key actually bills.
• Gemini Omni Flash is a different Google video model. Do not copy its duration string onto these rows.

Sibling compare (live): Kling vs Veo cost holds the 8-second Kling vs Veo dollar walkthrough. This page owns Veo only.

What a second buys

A billed second is not a free-form length. Official generate docs on 2026-09-18 pin what that second is allowed to be.

Lengths (verified on Google pages on the day checked)

• Veo 3.1 generate: video lengths 4, 6, 8 seconds. Standard card adds: reference image-to-video only supports 8 seconds.
• Vertex generate-video docs: Veo 3 models accept 4, 6, 8. Default 8. Veo 2 is 5 to 8, default 8. That is a different model.
• Gemini Veo generate guide: durationSeconds is "4", "6", "8". Must be “8” when using extension, reference images, or 1080p and 4K. Headline copy on that page still says Veo 3.1 generates 8-second videos (720p, 1080p, or 4k) with natively generated audio. 4s and 6s exist; they are not the 1080p / 4k row.

Do not write a 10-second or 15-second Veo 3.1 cell as one take. Ask for 10 seconds and you do not get a 10-second invoice. You get a refusal, then you generate 8 seconds (or two clips).

Gemini Omni Flash, on the same Vertex generate-video page, uses a duration string of integers 3 to 10 followed by "s" (example "10s"). That is not Veo 3.1. Keep the model id in the table.

Product fact (not Google marketing): lotgen’s Veo adapter accepts 4, 6, 8 and refuses 5, 7, and 9 even when an error string says “between 4 and 8, inclusive”. A page that bills “5 seconds of Veo 3.1” is pricing a length the predict path will reject.

Resolution and frame rate

Fast 4k is on both pricing pages ($0.30/s with audio). The Vertex Fast generate card on on the day checked lists 720p and 1080p only. Gemini generate docs list 4k for Veo 3.1 and Fast. This page prices the Fast 4k pricing cell and labels the generate-card gap. Confirm the resolution your model id actually returns before you budget 4k Fast on Vertex.

Default resolution on the Gemini generate guide: 720p.

Audio

• Gemini pricing: “video with audio price (default).” No silent Veo 3.1 row.
• Gemini generate guide, model features: Audio is Always on for Veo 3.1, Veo 3.1 Fast, Veo 3.1 Lite, and Veo 3 / Veo 3 Fast.
• Vertex Veo 3.1 generate cards: Sound generation: Supported on Standard, Fast, and Lite. The same cards print Audio as Not supported under Modalities. That line is input audio (mic_off), not “the clip is silent.” Do not read it as a silent SKU.
• Vertex pricing is the page that actually splits silent vs with-audio. That is the next section.

Product fact: lotgen’s Veo adapter treats audio as not optional on 3.1 (generateAudio is refused; published rates are the with-audio row).

Other generate-card facts (Vertex, on the day checked)

• Maximum output videos per prompt: 4.
• Standard: extend videos supported; reference images (asset images) supported.
• Fast: extend supported; reference images supported.
• Lite: extend printed Supported on Vertex; Gemini generate guide says extension is not available for Veo 3.1 Lite. Do not budget Lite extension until the path you call agrees with itself.
• Lite Vertex card also prints video generation as “Text to video, image to video, from first and last frame Supported” and Image modality Not supported. That is an internal contradiction on the same card. Gemini Lite row: Text-to-Video and Image-to-Video, no Video-to-Video.

Frame rate on Gemini and Vertex generate docs: 24fps.

Effective cost per second by tier

Formula, so the page still works after Google changes a cell:

clip dollars        = list dollars per second × billed seconds
usable-shot dollars = clip dollars × attempts

list dollars per second is the row for that tier, resolution, and audio flag. billed seconds is 4, 6, 8 on Veo 3.1, not the length you wished for. attempts is how many successful takes you run before you keep one. 3× below is a modeling assumption, not a measured keep rate. No lotgen.ai observed retry rate is on this page.

Gemini API, audio default (official per-second list, 2026-09-18)

Worked Fast 720p 8-second row:

$0.10/s × 8 s = $0.80 per successful take

Worked Standard 1080p 8-second row (same dollars as 720p Standard):

$0.40/s × 8 s = $3.20 per successful take

4s and 6s at 1080p or 4k are not computed here. Gemini generate docs require duration “8” at those resolutions. A 4-second 1080p Fast cell of $0.48 would be invented.

4K, 8 seconds, Gemini audio default (computed from the official 4k cells): Standard $4.80, Fast $2.40, Lite not supported.

Vertex silent rows (official / 1 count, treated as $/s for clip math)

If one blog says Veo is $0.20/s and another says $0.40/s, they may both be quoting Google: one silent on Vertex, one with audio on Gemini. That is two official rows, not a reseller typo.

Gemini API does not give you the silent column. Vertex does, if your job is actually silent. Confirm on the invoice. lotgen’s 3.1 path bills the with-audio row.

Fast vs Standard vs Lite: when the cheaper row is actually cheaper

List $/s is the sticker. “Fast is cheaper” is true on the sticker and false on the bill when the cheap row cannot do the shot, or when you throw enough takes that Standard would have been fewer dollars.

1. The sticker (Gemini audio default, 8 seconds, computed)

Quality is not in these tables. A cheaper second is not a better take. If a client contracted “Veo 3.1” without a tier, get the tier in writing. The 8-second 720p gap between Fast and Standard is $2.40 per take.

2. Retries can erase the Fast discount. 3× and 5× are a model, not a keep rate.

Three Fast 720p takes ($2.40) still beat one Standard take ($3.20). Five Fast 720p takes ($4.00) cost more than one Standard take. If Fast’s keep rate is much worse than Standard’s, the cheap row loses. This page has no lotgen run to measure that. Do not treat 3× as a forecast.

3. Length cap: you cannot buy a 4-second 1080p take on the Gemini generate path. Gemini docs: 1080p and 4k only support 8s. Fast at 1080p is $0.12 × 8 = $0.96, not $0.12 × 4. Iterate at 720p / 4s Fast ($0.40) and up-res later only if you accept a second billed take. Vertex’s Veo 3.1 generate cards list 4, 6, 8 without repeating the 1080p-must-be-8s rule except on Standard reference image-to-video. Budget 8s at 1080p unless the path you call documents otherwise.

4. 4K is a different product on Lite, and maybe on Vertex Fast. Lite: 4k output not supported on Gemini pricing, Gemini generate docs, and the Vertex Lite card. Fast 4k: priced at $0.30/s with audio on Gemini and Vertex; not listed as an output resolution on the Vertex Fast generate card on the day checked. Standard 4k is on both pricing and the Vertex Standard card. If the brief is 4k, Lite is not a cheaper Standard. Fast 4k is a maybe on Vertex until the generate card and the price grid agree.

5. Extension, references, image input. Gemini: video extension is Veo 3.1 and Fast only, not Lite; extension is 720p; each extend is 7 seconds, up to 20 times; output can combine input plus extension up to 148 seconds. That is additional billed video, not free runtime. Vertex Lite card prints extend Supported and reference images Not supported. Gemini Lite has no Video-to-Video row. If the job is “extend this 8s hero,” Lite is not the cheaper Standard on the Gemini guide.

6. A 10s or 15s brief is two Veo clips, not a discount. Two 8-second Fast 720p takes are $1.60 billed, then you cut. That is a stitch, not a list price. Not a lotgen run.

Gemini API vs Vertex: when the pricing model differs

Same model family. Two official tables. The invoice follows the key you paste.

A Gemini key and a Vertex key are not interchangeable credit piles. A consumer Gemini path does not fund a Vertex project, and a Vertex quota does not show up as Gemini API spend.

If you iterate silent on Vertex and add audio later, Vertex Standard silent 8s is $1.60 against Gemini’s $3.20. If you iterate with audio because that is what you ship, use the with-audio rows or you will under-budget by 2× on Vertex Standard.

Older Veo 3 (not 3.1) still sits on the Vertex grid on the day checked: with-audio $0.40 at 720p/1080p, silent $0.20; Veo 3 Fast with-audio $0.10 / $0.12, silent $0.08 / $0.10; no 4k cell on that older Fast row. This page’s worked clips are Veo 3.1. Do not paste a Veo 3 cell into a 3.1 table.

What an 8-second 720p / 1080p clip costs on each tier

Assumptions, all labeled:

• Model: Veo 3.1, Gemini API audio default (Always on in generate docs).
• Length: 8 seconds (default on Vertex generate-video docs; required at 1080p on the Gemini generate guide).
• Dollars: official Gemini per-second list, 2026-09-18.
• 1 attempt and 3 attempts. 3 attempts is a model, not a keep rate.
• Vertex silent is a second table, not mixed into these cells.

Standard does not get cheaper when you pick 720p. Fast and Lite do. That is the resolution split the sticker hides.

Vertex silent, same 8 seconds (computed from / 1 count as $/s):

4-second 720p, Gemini audio, 1 take (computed): Lite $0.20, Fast $0.40, Standard $1.60. Use this row only when the job is 720p and the API accepts 4.

Twenty kept 8-second 720p shots at the 3-attempt model: Lite $24, Fast $48, Standard $192. Swap the attempt slider before you treat that as a forecast.

Sibling Kling credits (live): Kling AI pricing explained. Sibling compare (live): Kling vs Veo cost. The cost pillar across every key: how much AI video generation costs.

How to cap what you spend

On Veo you find out the row after you pick Standard / Fast / Lite, resolution, and a length the API will accept. The control that matches a budget is a dollar quote before the take runs, on the keys you actually pay.

lotgen.ai is built as that control plane: one place to generate across the platforms it actually keys, see a real-dollar quote, compare takes, share a review link, and drive it via MCP. Invite-only beta. Public seat copy on lotgen.ai/pricing on 2026-09-18: $99 Solo, Team $90 a seat from 3, $79 from 6. You still pay Google for seconds. lotgen does not resell Veo tokens.

Twelve key platforms: Google, BytePlus ARK, fal, evolink, Kling, OpenAI, MiniMax, Black Forest Labs, Alibaba, ElevenLabs, xAI and Runway. Luma and Sora stay models on fal/evolink. BytePlus ARK is the console that issues the Seedance key, and the name to put on the invoice.

There is no lotgen.ai Veo run on this page. No side-by-side footage, no quote-UI screenshot, no measured keep rate. Same-prompt Kling vs Veo footage is a later brief. It needs lotgen run output and is not this page.

Live calculator, with this page’s defaults plugged in: 8 billed seconds, 1080p, audio on, 3 attempts, Google. The 3-attempt slider stays a model you set. Veo will not bill 10 seconds in one take.

Request an invitationSeats and pricingSee a dollar quote on your prompt.

Developers on a Gemini or Vertex key can use the same form: request an invitation.

Prices checked on 2026-09-18: how to re-check yourself

Refresh this page monthly, and within a few days of a Veo model bump.

1. Gemini per-second rows: Gemini API pricing, section “Veo 3.1”. Confirm Standard / Fast / Lite and 720p vs 1080p vs 4k. Google’s page was last modified 2026-09-16 when checked. Confirm audio still labeled default, and “charged only if successfully generated.”
2. Vertex silent vs audio, and / 1 count: Vertex generative AI pricing. Confirm the with-audio dollars still match Gemini, and whether “1 count” is still unexplained.
3. Lengths and 1080p/4k must-be-8s: Gemini Veo generate guide and Veo 3.1 generate. Confirm 4, 6, 8. Do not copy Omni Flash "10s" onto a Veo row. Re-read Vertex generate-video for the Veo 3 default of 8.
4. Fast 4k on Vertex: if the Fast generate card still omits 4K while the price grid lists it, treat that as a gap in the docs, not as a Fast 4k Vertex take you can buy.
5. Lite extension / image input: Gemini guide vs Vertex Lite card still disagreed when checked. Budget the path you call.
6. If you generate through fal, read fal Veo 3.1 the same day. Do not paste fal Fast $0.15/s over Google Fast $0.10 / $0.12, and do not bill a 5-second Veo 3.1 predict clip.
7. Recompute: $/s × billed seconds × attempts. If billed seconds are not 4, 6, 8 on Veo 3.1, stop and check the model id.

Related on this site: Kling vs Veo cost · Kling pricing explained · how much AI video generation costs · AI UGC ads cost · AI video cost calculator · how a render is quoted before it runs · seats and pricing · request an invitation

The buying guide for ads is best AI video generator for ads. A same-prompt Kling vs Veo compare with real footage is coming.

Request an invitationSeats and pricingA dollar quote on your own keys before a take runs.

FAQ