Veo R2V is 8 seconds at 128 credits

VEO 3.1 Reference to Video is 8 seconds, 128 credits silent or 256 with native audio, up to three reference images. That sticker is the price of an eight-second Veo clip with the reference stack attached. Not 64. Not 15. Not a 4-second draft. Eight seconds, 128 credits, image required. If the kit is the brief and the cinematic has to be Veo, you are on this row. If you wanted a pick between R2V and I2V, that is a different URL.

This is not Veo reference-to-video vs I2V. That page is the door: stack versus first frame. This page is the meter: 128 credits, 8 seconds only, three stills, native audio. The description on the card: generate an 8s cinematic clip whose subjects match up to 3 reference images, with native audio. No other duration is listed. No max output resolution is listed on this slug. Do not invent 4K here because Veo 3.1 T2V is 4K at 64 credits. Different card.

One hundred twenty-eight credits is the catalog sticker. Do not read it against T2V’s 64 as a “reference surcharge of 64.” Per second the two rows bill the same, 16 credits silent and 32 with sound; R2V only comes as eight seconds. You are not paying T2V plus a bolt-on. You are paying the R2V row.

Eight seconds is the whole menu

Duration is 8s. There is no 4s look-dev on this slug. There is no 6s. There is no 15s hold. Writing “continue for thirty seconds” into the prompt still returns eight. If the job is a half-minute identity hold, this is the wrong family. Grok Imagine Video starts at 24 credits, 720p, up to 30 seconds, T2V — no stack. Seedance 2.0 Fast Reference to Video is 55 credits (was 78), 720p, 4–15 seconds, native audio. Kling O3 Standard Reference to Video is 45 credits, 3–15 seconds, native audio. Those are other meters. Veo R2V is eight.

Plan the shot as eight seconds or change rows. A 128-credit take you will cut to three seconds is a legal edit and an expensive habit. A 128-credit take you will freeze-frame to twenty seconds is not a 20-second Veo file.

Native audio is on. Write the room. Dialogue, or none; footsteps, or none; music, or none. A silent prompt is not a mute plate. If you needed mute 20-credit motion, you wanted Wan 2.6 Image to Video Flash, and that row is I2V, not a stack.

Requires an image. Text-only will not run. If you have only a sentence, you are on Veo 3.1 T2V from 64 credits, 4 / 6 / 8 seconds, 4K — cheaper at 4s, the same rate per second, and no kit. Do not put a random JPEG in the R2V slot to force the 128-credit row. A random JPEG is not a brief.

Three stills is the kit, not a moodboard

Up to three reference images. Front of the spokesperson, three-quarter of the SKU, one clean plate of the room — then stop. A fourth near-duplicate is not on the card. Do not invent a 9-image Veo stack from a Symphony help page. TikTok Symphony’s Seedance 2.5 path went to 50 references (ppc.land, 9 August 2026). That is Symphony. This is Veo R2V at three.

Hygiene: same crop family, same light, no mixing a studio packshot with a phone grab. Mixed refs are how you get a person who is in none of the photos. Register the set once so clip nine pulls the same keys as clip one. Clip nine is nine times 128 if you rerun the cinematic every time. Spend cheap stills first. Re-rolling 128 credits because the label is the wrong colour is how a reference model becomes a sketchbook.

The prompt then names action and camera, not hair colour. The stills already answered who. Re-describing the subject in text is how the stack and the sentence fight.

Frame one is not your upload. The kitchen in the photo does not have to survive. The face and the SKU do. If the kitchen must survive, you wanted I2V or first-last, not this meter. VEO First Last Frame is 64 credits, 4K, 4 / 6 / 8 seconds, two stills as endpoints. Cheaper than 128 at its 4s floor, the same per second. Different contract. Endpoints are not a stack.

When 128 is the correct bill:

The product must appear in a place you did not photograph, and the cinematic has to be Veo’s eight with native audio.
The spokesperson must walk a new blocking without a new shoot, and you have a talent still you will not replace.
You have three approved stills and you would still send this shot to Veo if Seedance were free.

When 128 is the wrong bill:

You are still choosing among bottles. Approve the still on an image row. Grok Imagine Image Quality is 4 credits.
You need 15 seconds of the same kit. Kling Standard R2V at 45, or Seedance Fast R2V at 55, not Veo.
You need 4K dialogue without a kit. Veo T2V at 64.
You need mute look-dev. Flash I2V at 20.

The cost of insisting on Veo’s stack

Veo T2V, from 64 credits, is the sentence. Veo R2V, at 128, is the kit. The rate is the same; what you buy is the kit that stops Veo inventing a cousin of the bottle, and the eight-second length. It is worth it when the cousin would be a different SKU. It is not worth it when you do not have stills.

Seedance 2.0 full starts at 44 credits, 1080p, 4–15 seconds, T2V and R2V, native audio, and bills 25 credits a second at HD. That is a longer menu and a different family. Fast R2V is 55 at 720p, 14 credits a second. If 720p and 4–15s will ship, Fast is the draft/ref row and Veo’s 32 a second with sound is theatre. If the insertion order is “this must be Veo,” 128 is the honest sticker. Do not run Fast and grade it until it “feels like Veo.” It will feel like Seedance Fast.

Kling O3 Standard R2V is 45 for the first 5s and 9 a second after, so eight seconds is 72, 56 credits below Veo R2V’s silent 128, 3–15 seconds, no published 4K. If Veo is not the constraint, Kling is the kit door. If Veo is the constraint, Kling is a different movie.

Google Flow will run Veo. Flow is a surface. Optional visible watermark as of 14 August 2026 (Fastlane, 24 August 2026) does not change the 128-credit meter on this catalog. SynthID and C2PA stay embedded. Keep Veo. Do not live in Flow because the stack is Veo-shaped. Google Flow is a Veo surface, not a studio.

Sora 2 leftovers that were a character kit at 10–20 seconds are not an 8-second Veo R2V file. The Sora API shuts 24 September 2026 (OpenAI Help Centre, September 2026). Map the leftover to a family that has the length. Then, if an eight-second Veo cinematic is still the hero, pay 128 for that hero, not for the whole series.

Article 50 applies to photoreal people in the stills. 128 credits does not exempt a generated face. Label at first exposure for EU audiences since 2 August 2026. YouTube’s 27 May 2026 labelling pass overlays photoreal AI on Shorts. Labels do not refund the sticker.

Production order on 128

Lock three stills. Proof the pack at 100%. Proof the face. Then write an eight-second shot: one camera move, one verb, speech or none.

Do not use this row to discover whether the stills agree. Discover that on Seedance Fast R2V at 55, or on Kling Standard R2V at 45 if you need O3’s window. Promote the kit to Veo R2V when the cinematic is the ship file.

Captions after the eight. Native audio is not typesetting. The AI caption generator is the burn-in. The AI video generator is the launcher.

If you cannot name the three stills, you are not on this row. If you cannot name an eight-second shot, you are not on this row. If you can name both, 128 is the price of Veo’s eight.

Spend the 128 on an eight you already designed in Versely.

FAQ

Why is R2V 128 credits when Veo T2V starts at 64?

Different slug, different job, same rate. T2V is a sentence at 4 / 6 / 8s, 4K, from 64 credits for 4s. R2V is an 8-second kit match, 128 credits, up to three stills. An 8-second T2V clip costs the same 128. You are not being surcharged for the stack. You are on the reference row.

Can I draft Veo R2V at 4 seconds?

Not on this card. 8s only. Draft the kit on a cheaper R2V family, then run the eight at 128.

Is Veo R2V 4K?

The published R2V card does not list a max output resolution. Do not assume the T2V 4K token. If 4K is the insertion order for a sentence, use Veo 3.1 T2V at 64. If 4K first-last is the job, that card is 64 credits and two endpoints, not this stack.