# AI video models compared: Veo 3.1, Kling, Wan and Sora's exit

URL: https://infro.io/blog/ai-video-models-compared-2026
Author: INFRO Editorial (https://infro.io/blog/editorial)
Publisher: INFRO
Published: 2026-09-29
Last modified: 2026-09-29
Last reviewed: 2026-09-29 (next review due 2026-10-29)
Topic: Agents and multimodal systems
Format: Comparison

> Compare video models on what their vendors document — clip length, resolution, native audio, job and retention semantics, and per-second list rate — then test the two or three that fit on your own prompts. Sora 2 was shut down on 24 September 2026, which is the strongest argument for keeping the model choice in configuration.

Five days before this article was written, OpenAI shut down its Videos API. The deprecations page lists the Videos API, `sora-2` and `sora-2-pro` — every variant — as announced on 24 March 2026 and shut down on 24 September 2026, with no replacement named[1]. The Sora 2 model page now says the model and API were discontinued on that date[2].

That is the most useful fact in any video model comparison this year. Video models are young, their APIs change shape often, and a model you build a product around can leave with six months' notice. So this comparison sticks to what vendors document, gives you a way to choose for your own prompts, and treats "how hard is it to move" as a first-class criterion.

## What to compare, and what not to

Leaderboards for video are subjective, change monthly and rarely match a specific product's prompts. The facts that decide whether a model can serve a workload at all are more stable and are published by each vendor:

- **Clip length** per generation, and whether clips can be extended.
- **Resolution and aspect ratio**, and the combinations that are allowed together.
- **Audio**: generated natively, supplied by you, or absent.
- **Job semantics**: how you submit, poll or receive a callback, and the status values.
- **Output retention**: how long the finished file stays downloadable at the provider.
- **List rate** per second of output, and what moves it.

Everything below comes from the vendor's own page, opened on 29 September 2026. Where a page did not state something, the table says so rather than guessing.

## The documented facts, side by side

[Diagram: Documented facts from each vendor's own documentation, 2026-09-29. "Not stated" means the page opened for this article did not say, not that the model lacks the feature.]

**Google Veo 3.1.** The Gemini API offers three preview models: `veo-3.1-generate-preview`, `veo-3.1-fast-generate-preview` and `veo-3.1-lite-generate-preview`. Clips are 4, 6 or 8 seconds; 1080p, 4K, reference images and extension all require 8 seconds. Veo 3.1 and Fast render 720p, 1080p and 4K; Lite stops at 1080p. All three generate audio natively, aspect ratios are 16:9 and 9:16, and generated videos are kept on Google's servers for two days. Google documents request latency from 11 seconds to 6 minutes at peak, and every video carries a SynthID watermark[3].

**Kling VIDEO 3.0 series.** Kling's own guide says the 3.0 series generates from 3 to 15 seconds, with native audio covering dialogue, ambience, effects and lip movement, native 4K output, and multi-shot generation that sets framing and camera movement per shot within one request[4].

**Alibaba Wan 2.7.** The text-to-video reference lists `wan2.7-t2v` (with a dated snapshot), 2 to 15 second clips defaulting to 5, 720P or 1080P, and five aspect ratios. Without an `audio_url` the model adds background music or effects; with one, it uses your WAV or MP3. The task id and the video URL are both valid for 24 hours only[5].

**MiniMax H3.** MiniMax's current video guide lists MiniMax H3 (768P or 2K, 4–15 seconds) and H3 Max (480P or 768P, 5–15 seconds), with text-to-video, first- and last-frame image-to-video, and generation from reference images, video or audio[6].

**xAI Grok Imagine video 1.5.** `grok-imagine-video-1.5` generates up to 15 seconds; you start a request, poll `/v1/videos/{request_id}`, and stop at `done`, `failed` or `expired`[7]. xAI's model page lists it at $0.080 per second[8].

**ByteDance Seedance 2.5.** BytePlus ModelArk documents Dreamina Seedance 2.5 behind a task API with create, retrieve, list and cancel-or-delete operations[9]. The model's own tutorial did not load for this article, so its clip limits are not listed here.

## List rates per second

Video is billed per second of output, and resolution usually moves the rate. The chart shows only rows in INFRO's model catalog whose vendor rate carries a checked source.

[Diagram: Vendor list rates per second of output at 720p, from Google's Gemini API pricing page and Runway's API pricing page. A negotiated rate on your own account replaces these.]

Google lists Veo 3.1 at $0.40 a second at 720p and 1080p and $0.60 at 4K; Veo 3.1 Fast at $0.10, $0.12 and $0.30; Veo 3.1 Lite at $0.05 and $0.08[10]. Runway lists Gen-4.5 at $0.12 a second and Gen-4 Turbo at $0.05[11]. For Sora 2 the model page still shows $0.10 a second at 720p[2], but that rate no longer buys anything.

> **Example: What one clip costs at list rate**
> An 8-second 1080p Veo 3.1 clip at $0.40 a second is $3.20 at list rate. The same clip on Veo 3.1 Fast at 1080p ($0.12) is $0.96, and on Lite ($0.08) is $0.64. If a product generates three drafts for every clip a user keeps, the cost per kept clip is three times those figures — which is why cost per accepted result, not per render, is the number to compare.

Two practical notes. Per-second rates multiply by the seconds actually rendered, so a model whose minimum clip is longer than you need charges for footage you discard. And a rate on a vendor page is what an account without a negotiated agreement pays; if you have a contract, your own rate is the one that matters.

## Asynchronous jobs are the real integration work

Every video API above is asynchronous. The shape is the same — submit, receive an id, poll or receive a callback, download — but the details differ in ways that break naive code:

- **Status vocabulary.** OpenAI's retired API used `queued`, `in_progress`, `completed` and `failed`[12]. Wan uses `PENDING`, `RUNNING`, `SUCCEEDED`, `FAILED`, and `UNKNOWN` after expiry[5]. xAI uses `done`, `failed` and `expired`[7]. A client written for one maps poorly onto another.
- **Retention.** Google keeps Veo output for two days[3]; Wan's URL lasts 24 hours[5]; OpenAI's download URLs lasted at most one hour[12]. Treat every provider URL as a handoff and copy the file to your own storage.
- **Latency.** Google's documented range runs from seconds to minutes[3], and OpenAI noted that a single render could take several minutes, longer for higher resolutions and durations[12]. Nothing in a video path belongs inside a user-facing HTTP request.
- **Content rules.** OpenAI's guide listed rejections for real people, copyrighted characters and music, and input images with human faces[12]; Google restricts person generation settings in some regions[3]. Test your actual prompt categories, including refusals.

## How to choose for your workload

1. **Filter on hard constraints.** Required clip length, resolution, aspect ratio, audio, and regions you may process in. This usually removes half the list.
2. **Write a fixed prompt set** from real product requests: 20 to 50 prompts, including the ones that must be refused.
3. **Render each prompt on the remaining two or three models** with the same duration and resolution, and have the people who own the output rate them blind.
4. **Compute cost per accepted clip**: list or negotiated rate × seconds × renders per accepted result.
5. **Record the job behaviour you observed**: time to completion, failure rate, and how failures were reported.
6. **Keep a second model qualified.** A shutdown with no named replacement is exactly the case where an already-tested alternative is worth having.

## Running video jobs through INFRO

INFRO documents one asynchronous job API for video: `POST /v1/videos` returns `202` with a job in `queued`, and the job moves through `running` to `succeeded`, `failed` or `canceled`. The submit is checked against your budgets using the requested duration; INFRO then submits the render to the first of your connections that accepts it, trying the next target on your route if a submit fails, and checks the provider for status about 15 seconds after submitting and then every one to five minutes. Once a provider has accepted a render it is never moved, and a render unfinished after 60 status checks fails with `408 request_timeout`. See [video generation](https://infro.io/docs/modalities/video) and [async jobs](https://infro.io/docs/api/jobs).

[Diagram: A video job through INFRO. The provider is your own connection; failover happens only while submitting, and the output link lasts 24 hours.]

Three details matter in production. Send an `Idempotency-Key` on the submit: a retry with the same key returns the original `202` instead of starting, and paying for, a second render. Prefer [webhooks](https://infro.io/docs/api/webhooks) over polling — deliveries to endpoints configured in the console are signed with an HMAC you verify, retried on a widening schedule for about 16 hours, and may arrive more than once, so deduplicate on the event `id`. And copy the output: `output.url` expires 24 hours after completion, when INFRO deletes its copy.

A succeeded job carries `usage.seconds` and `usage.cost`, an estimate of what your provider will bill, with `price_source` naming the rate used; `cost` is `null` for an unpriced model, and INFRO records no cost for failed or canceled jobs. INFRO charges nothing per second; its subscription is separate from what your provider bills.

> **Warning: Video serving status today**
> INFRO's documentation states that none of the provider kinds a customer can connect today serves video, so `POST /v1/videos` answers `503 no_available_provider` until one does. The job, webhook and idempotency contract above is documented and stable; which vendors you can reach through it depends on the provider connections that ship. Check [providers](https://infro.io/providers) before planning a video workload on it.

Keeping the model name in a [route](https://infro.io/docs/routes) rather than in code is the lesson from Sora 2 either way: when a vendor retires a model, you change the route's targets, not every service that renders video.

## Questions

### What replaced Sora 2 in the OpenAI API?

OpenAI's deprecations page lists the Videos API, sora-2 and sora-2-pro as shut down on 24 September 2026 with no replacement named. Applications that called it need a video model from another vendor.

### Do AI video APIs charge for failed renders?

It depends on the provider, and the vendor pages opened for this article do not all state it. Check your provider's terms; a gateway can record that a job failed, but only your provider's invoice says what was billed.

## Sources

1. Deprecations — OpenAI. https://developers.openai.com/api/docs/deprecations (accessed 2026-09-29)
2. Sora 2 model page — OpenAI. https://developers.openai.com/api/docs/models/sora-2 (accessed 2026-09-29)
3. Generate videos with Veo 3.1 in the Gemini API — Google. https://ai.google.dev/gemini-api/docs/veo (accessed 2026-09-29)
4. Kling VIDEO 3.0 cost guide: 4K, Omni audio and multi-shot — Kuaishou (Kling AI). https://kling.ai/blog/kling-video-3-0-credit-cost-guide (accessed 2026-09-29)
5. Wan text-to-video API reference — Alibaba Cloud Model Studio. https://www.alibabacloud.com/help/en/model-studio/text-to-video-api-reference (accessed 2026-09-29)
6. Video generation guide — MiniMax. https://platform.minimax.io/docs/guides/video-generation (accessed 2026-09-29)
7. Video generation guide — xAI. https://docs.x.ai/docs/guides/video-generations (accessed 2026-09-29)
8. Models and pricing — xAI. https://docs.x.ai/docs/models (accessed 2026-09-29)
9. Create a video generation task — BytePlus ModelArk. https://docs.byteplus.com/en/docs/ModelArk/1520757 (accessed 2026-09-29)
10. Gemini Developer API pricing — Google. https://ai.google.dev/gemini-api/docs/pricing (accessed 2026-09-29)
11. API pricing — Runway. https://docs.dev.runwayml.com/guides/pricing/ (accessed 2026-09-24)
12. Video generation with Sora — OpenAI. https://developers.openai.com/api/docs/guides/video-generation (accessed 2026-09-29)