Sora vs Veo cost

The short answer

Both figures are list price for one clip at one stated duration and resolution. Neither is a lotgen quote. Neither is a measured keep rate. There is no footage on this page, and there is no winner column.

$0.085 times 8 is $0.680. evolink’s 8 second row says $0.678. The worked Sora cell uses the printed per-video charge, not the headline times the duration.

What the keyword is asking

sora vs veo cost is a list-price question. It asks what a Sora clip and a Veo clip cost before anyone runs a lot. It is not a beauty contest, and this page does not pick a look.

A same-prompt quality comparison is a different job. It needs the same prompt, a real run, and a keep rate. That is not this page.

The two columns above share a duration and a resolution that both routes actually sell: 8 seconds at 720p. Veo also sells 4 and 6 seconds at 720p. It does not sell an arbitrary 15 second take. At 1080p and 4k, the Gemini guide says the length must be 8 seconds. The evolink Sora route sells 4, 8, and 12 seconds, at 720p only.

A lower printed list price is not a verdict about which clip you keep.

Veo list price

One worked row, from Google’s own pages on 2026-09-21.

Gemini API, paid tier, Veo 3.1 Standard, video with audio (the default row): $0.40 per second at 720p and at 1080p. At 8 seconds and 720p, list price is $0.40 x 8 = $3.20. Free tier for this model is marked not available. Google says you are charged only if the video is successfully generated.

The same Gemini table, same day, not worked out here: Fast is $0.10 per second at 720p, $0.12 at 1080p, and $0.30 at 4k. Lite is $0.05 at 720p and $0.08 at 1080p, and the row says 4k output is not supported. Standard with audio at 4k is $0.60 per second. What Standard, Fast, and Lite mean lives on Veo pricing explained.

Source: Gemini API pricing, model ids veo-3.1-generate-preview, veo-3.1-fast-generate-preview, and veo-3.1-lite-generate-preview. Duration and aspect: Generate videos with Veo 3.1. durationSeconds is “4”, “6”, or “8”, and must be “8” for extension, reference images, or 1080p and 4k.

Vertex still has its own table. Vertex generative AI pricing lists the same with-audio dollars, written as “$0.40 / 1 count” for Veo 3.1 at 720p and 1080p, and “$0.60 / 1 count” at 4k. It also prints a video-only table (no audio): Veo 3.1 at 720p and 1080p is “$0.20 / 1 count”, and 4k is “$0.40 / 1 count”. A Vertex “count” is not multiplied by 8 seconds here, because the cell does not say second. The worked row above is the Gemini per-second price, not the Vertex count.

Vertex model ids on the Veo 3.1 card are veo-3.1-generate-001, veo-3.1-fast-generate-001, and veo-3.1-lite-generate-001 (Lite is marked Preview). Those are not the Gemini -preview ids. The card says sound generation is supported, audio as an input is not supported, pay-as-you-go is not supported, and fixed quota is supported. That is a different sentence from the Gemini paid-tier price. Both pages were live on 2026-09-21. They are two bills, not one.

Sora list price

Sora on lotgen is a model on a route, not an OpenAI key. The worked cell in the table is the route the product catalog actually lists. The other two prices below are labeled as what they are.

What OpenAI itself publishes

OpenAI does publish a per-second API price. It is not a blank, and it is not a ChatGPT subscription converted into dollars per second.

API pricing, video generation, standard, read 2026-09-21:

Batch on the same page is half of each of those: $0.05, $0.15, $0.25, and $0.35 per second. The model pages match the standard column: sora-2 at $0.10 per second for those 720p sizes, and sora-2-pro at $0.30, $0.50, and $0.70. Both model pages say the output is video and audio. The price table has no separate audio line.

A worked OpenAI list price, not the lotgen route: sora-2 at 720p for 8 seconds is $0.10 x 8 = $0.80 on the standard table, or $0.40 on the batch table. That $0.80 is OpenAI’s API list. It is not what the OpenAI key on lotgen’s request form buys.

Deprecations, read the same day: the Videos API, sora-2, sora-2-pro, and the named snapshots shut down on 2026-09-24. The recommended-replacement column is empty. No successor is named. The video guide still documents that API, including generations up to 20 seconds, and says each extension can add up to 20 seconds, up to a total of 120 seconds. The guide’s banner names the same shutdown date.

OpenAI’s consumer pricing page redirects to chatgpt.com/pricing. Plus and Pro are subscriptions, and a subscription does not convert into a per-second rate, so neither is quoted here.

fal and evolink, labeled as routes

evolink is the route lotgen’s catalog lists. On 2026-09-21 the product catalog has one Sora row: evolink sora-2-preview. There is no fal Sora row.

evolink’s Sora 2 page says it routes sora-2-preview only, and that Sora 2 Pro, Beta, and Beta Max routes were retired. The page prints:

• $0.085 (about 5.76 credits) per second of video
• 4 seconds: $0.339 per video (23.04 credits)
• 8 seconds: $0.678 per video (46.08 credits)
• 12 seconds: $1.017 per video (69.12 credits)

The worked cell uses the 8 second printed charge, $0.678. That is evolink’s list price for that row. It is not an OpenAI API rate, and it is not a lotgen quote. The page says the route is billed per second and also prints those three per-video charges. They disagree by two tenths of a cent if you multiply the $0.085 line, so this page does not treat $0.085 as a rate you can multiply.

fal publishes a Sora endpoint. lotgen’s catalog does not list it. fal-ai/sora-2/text-to-video was still up on 2026-09-21 and says: “The pricing is $0.1/s for Sora 2.” A worked fal list price at the same 8 seconds and 720p is $0.10 x 8 = $0.80. That is fal’s price for fal’s endpoint. It matches OpenAI’s standard sora-2 720p number as of that day. Matching numbers are not the same bill. The endpoint’s published schema no longer carries a price field, so the $0.80 rests on that sentence on the page.

The same fal page says the endpoint will shut down on September 24, 2026, that the endpoint is deprecated, and that the model is no longer supported. The page is still up, and it still prints the $0.10 per second sentence. Because the catalog has no fal Sora row, lotgen does not quote this $0.80.

fal’s schema for that endpoint: resolution is fixed at 720p, aspect ratio is 9:16 or 16:9, duration is 4, 8, 12, 16, or 20 seconds. The page describes clips with audio. It does not print a separate audio price.

No other Sora rate is filled in. If evolink retires sora-2-preview and nothing replaces it in the catalog, the Sora column goes with it.

What the official docs actually limit

Only the limits below are from a page linked here. There are no quality scores. Checked on 2026-09-21.

evolink sora-2-preview (page): durations 4, 8, and 12 seconds. Sizes 1280x720 and 720x1280. No 1:1 on the page. The page describes synchronized dialogue and sound effects and calls the model a video-and-audio model. Max length sold: 12 seconds.

fal fal-ai/sora-2/text-to-video (page, schema): 720p only. Aspect 9:16 or 16:9. Duration 4, 8, 12, 16, or 20 seconds. No 1:1 in the schema. Described as clips with audio. The page says this endpoint shuts down on September 24, 2026. Not a lotgen catalog row as of 2026-09-21.

OpenAI’s own Videos API (pricing, guide): priced sizes are the portrait and landscape pairs in the table above. No square size is in that table. The guide says generations up to 20 seconds, and extensions up to a total of 120 seconds. Model pages list video and audio as outputs. This API is scheduled to shut down on 2026-09-24, with no replacement named. The guide mentions 480p only inside a latency sentence. The pricing table has no 480p row, so there is no 480p figure here.

Veo, Gemini (guide): 4, 6, or 8 seconds, and 8 seconds is required at 1080p, at 4k, for reference images, and for extension. Aspect 16:9 or 9:16. No 1:1. Native audio on the default price. One generation maxes at 8 seconds. Extension, on that guide, runs up to 148 seconds, and video extension on that guide is limited to 720p.

Veo, Vertex (card): video lengths 4, 6, or 8 seconds. Reference image to video on the Standard card supports 8 seconds only. Output resolutions on Standard: 720p, 1080p, and 4k. Fast and Lite cards list 720p and 1080p. Aspect 9:16 and 16:9. Sound generation supported. Audio as an input is not supported. The card says pay-as-you-go is not supported and fixed quota is supported. Read that next to the Gemini per-second paid tier, not instead of it.

Sora is not its own key on lotgen. The OpenAI row on the request form is a key for GPT Image 2 stills and GPT text drafts. It is not a Sora key. The twelve media keys are Google, BytePlus ARK, fal, evolink, Kling, OpenAI, MiniMax, Black Forest Labs, Alibaba, ElevenLabs, xAI and Runway. The form’s thirteenth row is Anthropic, and it is text only. Luma and Sora are models, not keys. ElevenLabs is voice.

What this page will not decide

This page will not tell you which clip looks better, what share of takes you keep, or what a same-prompt pair costs after a real run. Those need a lot you can watch. There is no footage here, and no stills are invented to stand in for one.

kling vs veo vs sora is a three-way quality question. A same-prompt footage comparison of those models does not exist yet, so there is nothing here to answer it with.

A different pair, Kling versus Veo in dollars, is already a page: Kling vs Veo cost. Kling’s own dollar crosswalk is Kling pricing explained. The wider per-second list is how much AI video generation costs. The cost calculator prices a spec you type. Veo’s Standard, Fast, and Lite grid lives on Veo pricing explained.

It does not pick a best model for ads either, and it does not rank generators.

Request access

lotgen is where the keys you already have can be quoted before you commit to a lot. The seat is not the clip. The labs still bill the clip.

The invitation form lists thirteen platforms. Twelve of them are media keys. Anthropic is the thirteenth, and it is copy, text only. Luma and Sora are models on fal or evolink, not keys. ElevenLabs is voice. The OpenAI key is not a Sora key.

Request an invitationSeats and pricing

The quote you get there is the keys on the account. It is not the list-price table above, and it is not a keep rate.

Team aside

FAQ