Applied artificial intelligence, neural systems, and smart agents.

Fusing generative AI, multi-agent frameworks, and high-throughput data processing to automate business workflows.

Version

Veo 3.1

Veo 3.1 generates four, six or eight second clips, with audio permanently enabled and no option to disable it. 4K support is present, but restricted to the eight-second clip, and reference capacity reaches up to three images.

  • four, six or eight seconds
  • audio always on
  • three reference images

preview Provider: Google

Last checked: This is a reference page. It is re-checked against the vendor sources and updated when a new version ships. Change log

Eight seconds, and the three operating conditions attached to it

Google documentation for this unit records three clip lengths: four, six and eight seconds. The eight-second tier is conditional, and the same documentation states the condition: eight seconds is available only when the output is 1080p or 4k, or when reference images are used. Resolution follows the same conditional structure: 1080p and 4k are active only on that eight-second clip.

So this unit 4K figure is not equivalent to Kling 4K figure. Kling provisions 4K at any clip length; Veo provisions it only at its own length ceiling. Both are recorded in full in the resolution column of our table, and this distinction belongs on this page.

Extension: the actual route to a longer video

Veo extends a video it generated earlier: seven seconds per operation, up to twenty times. On specification alone, that constitutes a long video. In operation there is a constraint the documentation states plainly: during extension, the output is 720p only. So either the long video or the 4K output is available — not both simultaneously.

Three-image reference capacity

The reference ceiling is three images, and the documentation notes this capability is active on the 3.1-series units only. Three, against fifty for Seedance 2.5 and seven for Kling, is the lowest figure in our table. It is sufficient for the operation "place this character in this scene." It is not sufficient to construct a consistent world with several characters and several products.

Why it remains a widely deployed choice despite these constraints

Because it is the only unit in this table reachable through a public, documented API with an ordinary key, and its technical documentation is readable without an account. For a team integrating video generation inside its own product, that single factor typically outweighs a clip twice as long.

Specifications

Every number here comes from the vendor's own page, with that page linked beside it.

Unit identifier
veo-3.1-generate-preview Source
Input
text, image Source
Output
video, audio Source
Deployment surface
Gemini API, Runway API Source
Longest clip in one pass
8 seconds Source
Resolution ceiling
4K Source
Vertical resolution
2,160 lines of vertical resolution Source
Native audio
yes Source
Reference input capacity
3 reference inputs Source

Using it from Iran

This column carries a date and says how it was checked, because most listicles guess it. Where we have only read a vendor policy page, the note below says exactly that.

Reachable
blocked
Payment
no working route
Free tier
no
Measured on

How we checked: Google supported-countries list was checked and read directly today: the entry after Indonesia jumps straight to Iraq, and Iran is not recorded on this list. This finding comes from reading a document, not from a network measurement; a direct test from inside Iran has not yet been run.

Documented strengths

  • Audio is permanently enabled, eliminating a separate step to add it
  • 4K support on the eight-second clip, matched in this table only by Kling
  • Seven-second extensions, up to twenty of them, on a video it generated itself
  • Open, readable technical documentation, on Google public API

Room for improvement and technical considerations

  • Eight seconds is the ceiling, and the shortest length in this table.
  • 4K and 1080p are active only on the eight-second clip, not at four or six seconds.
  • An extended output is 720p only, so a long video and 4K cannot be combined.
  • Three-image reference capacity is the lowest figure in our table.
  • Iran is not recorded on Google supported-countries list, and no official route exists.

Technical verdict

For integrating video generation inside software under development, this is the most straightforward integration path, and its technical documentation is readable in half an hour. For a workload where the finished output matters more than the route to producing it, the eight-second ceiling and three-image reference capacity become constraining quickly.

Frequently raised questions, with documented answers

How long a video does Veo 3.1 generate

Four, six or eight seconds. The eight-second tier is available only when the output is 1080p or 4k, or when reference images are used.

Can Veo produce a longer video

Through the extension operation, yes: seven seconds per operation, up to twenty times. But the output of the extension path is 720p only, so a long video is not produced in 4K.

Is an access route to Veo available from Iran

No official route exists. Iran is not recorded on Google supported-countries list; that list was checked directly today. This platform does not sell accounts and does not propose a route around this restriction.

Sources

  1. Google, Veo documentation in the Gemini APIvendor sourceai.google.devread on 12 August 2026
  2. Google, Gemini API model listvendor sourceai.google.devread on 12 August 2026
  3. Where you can use the Gemini web appvendor sourcesupport.google.comread on 12 August 2026
  4. Runway API pricing, per second of outputvendor sourcedocs.dev.runwayml.comread on 12 August 2026