Veo 3.1: modes, aspect ratios and credit pricing

veo-3.1-generate-preview is Google’s flagship video model and Playcut’s default. It generates 24 fps MP4 with native synchronized audio — dialogue, sound effects and ambience — in a single pass, and it is the only model in the Playcut catalog that supports all five video modes (text-to-video, animate, reference, interpolation, extension) and the only one that reaches 4K. It is also the most expensive video model we sell: 160 credits per second at 720p and 1080p, 240 credits per second at 4K. Pick Veo 3.1 when motion realism, audio, or 4K matters. For volume work, the Grok and DashScope models are 3–8× cheaper per second.

Specs at a glance


Model id	`veo-3.1-generate-preview`
Provider	Google (Gemini API)
Modes	text-to-video, image-to-video, reference-to-video, interpolation, extension
Aspect ratios	`16:9`, `9:16` — nothing else
Resolutions	`720p`, `1080p`, `4k` (lowercase)
Duration	`4`, `6`, or `8` seconds — a fixed enum, not a range
Max reference images	3 (reference mode)
Output	MP4, 24 fps, 1 video per request, audio always on
Credits	160 cr/s @ 720p and 1080p · 240 cr/s @ 4k

Where the 8-second rule applies

Duration must be 8 seconds whenever any of the following is true:

resolution is 1080p or 4k
reference images are used
the request is an extension

At 720p with no references you may choose 4, 6, or 8 seconds. There is no 4-second 4K clip.

Modes

Mode	Playcut tool	Input	Refs	Constraints
Text to video	`generate-video-from-text`	prompt	—	4 / 6 / 8 s
Animate (image to video)	`generate-video-from-image`	1 source image + prompt	1 image	source is letterboxed to the target frame
Reference to video	`generate-video-from-reference`	prompt + reference images	up to 3	16:9 only, 8 s only
Interpolation	`generate-video-interpolation`	start image + end image	2 (first/last frame)	always 8 s
Extension	`generate-video-extension`	a previously generated Veo video	1 video	720p only, 8 s; adds 7 s per call

Video editing (source video + prompt) is not a Veo capability on Playcut — that mode belongs to Wan 2.7.

Extension limits

Each extension call adds 7 seconds of new footage, up to 20 times, on a source video of at most 141 seconds. The source must be a video Veo itself generated; an uploaded MP4 cannot be extended. Google stores generated videos for 2 days, so a clip older than that can no longer be extended.

Pricing

Credits are held at submission — durationSeconds × credits-per-second — and refunded in full if the task fails or is cancelled.

Resolution	Credits / second	4 s clip	6 s clip	8 s clip
720p	160	640	960	1,280
1080p	160	—	—	1,280
4k	240	—	—	1,920

Dashes mark combinations the 8-second rule forbids. Google’s own list price for Veo 3.1 is $0.40/s at 720p and 1080p and $0.60/s at 4K (verified against the Gemini API pricing page on 2026-07-21).

Playcut does not offer Veo 3.1 Fast or Veo 3.1 Lite.

Limitations & gotchas

Only two aspect ratios. 16:9 and 9:16. Square, 21:9, and every other ratio are rejected by Playcut’s pre-submission guard (since 2026-07-23) with an instant 400 — before any credits are held — on every mode including API and MCP calls. The same guard forces the 8-second duration at 1080p and 4K.
4k must be lowercase. "4K" is rejected.
The 8-second tax. Anything premium — 1080p, 4K, references, extension — forces 8 seconds, so the cheapest possible 4K clip costs 1,920 credits.
Reference mode is 16:9 only on Playcut, in addition to being 8 s only.
Extension is 720p only. You cannot extend a 1080p or 4K clip.
Animate letterboxes rather than crops. A source image whose ratio differs from the requested aspect ratio gets black bars baked into the video.
seed is not supported by Google’s SDK for video generation. We record it on the asset for provenance but never send it.
Audio is always on. There is no toggle to generate a silent clip.
Person generation is region-restricted. In the EU, UK, Switzerland and MENA, Google limits Veo 3 / 3.1 to adult-only person generation. Playcut does not currently adjust for this, so requests from those regions may be rejected upstream.
Latency is long and bursty. Google quotes 11 seconds minimum and up to 6 minutes during peak hours.
Prompts are capped at 1,024 tokens by the provider. We do not truncate for you.
Every output carries a SynthID watermark.

How to generate

Veo 3.1 is the default video model, so you only need to pass a model id if you want a different one. See the Quickstart for session creation, polling and download.

curl -X POST https://api.playcut.ai/api/v1/tools/generate-video-from-text/execute \
  -H "Authorization: Bearer $PLAYCUT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workerId": "65f3a1b2c3d4e5f600000001",
    "prompt": "A lighthouse in a storm, waves crashing, cinematic",
    "aspectRatio": "16:9",
    "resolution": "720p",
    "durationSeconds": 8
  }'

curl -X POST https://api.playcut.ai/api/v1/tools/generate-video-from-image/execute \
  -H "Authorization: Bearer $PLAYCUT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workerId": "65f3a1b2c3d4e5f600000001",
    "prompt": "Slow push in, hair moving in the wind",
    "imageAssetId": "65f3a1b2c3d4e5f600000003",
    "aspectRatio": "9:16",
    "resolution": "720p",
    "durationSeconds": 6
  }'

curl -X POST https://api.playcut.ai/api/v1/tools/generate-video-from-reference/execute \
  -H "Authorization: Bearer $PLAYCUT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workerId": "65f3a1b2c3d4e5f600000001",
    "prompt": "The woman in the red coat walks through a night market",
    "referenceImageAssetIds": ["65f3a1b2c3d4e5f600000003"],
    "aspectRatio": "16:9",
    "resolution": "720p",
    "durationSeconds": 8
  }'

The same tool names are exposed over MCP — generate-video-from-text, generate-video-from-image, generate-video-from-reference, generate-video-interpolation, generate-video-extension. See the MCP install guide.

All models The full Playcut catalog — three image models, seven video models, music and voice.

Seedance 2.0 1080p at 52 cr/s and 15-second clips — the cheap alternative when 4K is not required.

Grok Imagine Video The budget video tier at 20 cr/s, 720p only, with reference-to-video and extension.

Wan 2.7 The only model that lip-syncs to driving audio and the only one that edits an existing video.

Aspect ratios by model Which ratios every Playcut image and video model accepts.

Resolutions and pricing Credit cost per second and per image across the whole catalog.