Testing out creating videos in Google Gemini 3

This page has 0 threads | Add post

Blogs

Many people around the world enjoy and benefit from our blogs. If you're one of them - and can afford to - please make a small donation to help keep this and future blogs, exercises, skills assessment tests, tips, tutorials and videos freely available to everyone!

Some other pages relevant to this blog include:

AI (Artificial Intelligence) training courses

You can also book hourly online consultancy for your time zone with one of our 7 expert trainers!

Testing out creating videos in Google Gemini 3
Having had disappointing experiences with Sora before, we thought we'd see if Google Gemini's Veo text-to-video generation was better.

Posted by Andy Brown on 29 January 2026

I have no doubt that within 5 years AI tools will be able to generate videos (from text instructions) which are indistinguishable from real camera footage - but that day has not yet arrived.

I wanted to include OpenAI's Sora video generation tool in this review, but the first version was awful and despite playing about with VPNs and Discord servers I can not get access to Sora 2 to try it.

To choose to generate a video in Google's Gemini 3 Pro AI tool (I'm using the paid version, as the name suggests) click on the relevant button:

Click on this button to create a video in Gemini.

I had scripted the prompt for 4 different videos to test different aspects of Gemini, but in the end only got to try this one (together with two prompts I tried earlier in the week):

"A typical montage scene at a self-scanning kiosk in a UK supermarket, except that the customers are all farmyard animals."

Judge the results for yourself (they took about 1 minute 10 secods to appear, which is pretty quick):

This is pretty impressive, even though the quality isn't yet anywhere good enough for commercial use. I had previously also tried this prompt:

"Create a video showing a close-up shot of George Orwell walking near his house in Barnhill in Jura."

Here's what this gave:

Although this looks nothing like George Orwell, it's a big improvement on previous videos I've generated from the same prompt in Sora.

And finally, I'd tried this whimsical prompt:

"The video begins with a soft focus shot of a couple in their 20s lying beside a pool in a holiday resort in the evening light. As the camera gradually pans in the man in the couple gradually morphs into a crocodile and the woman gradually changes into a shark, while the swimming pool gets hotter until by the end of the video it is boiling."

Which gave this slightly disturbing video:

All of which is quite impressive - certainly leaps and bounds better than Sora version 1. But my problem is that after generating a single video using Gemini today, I got this message:

The disappointing message following only my second video request.

To me this makes this tool virtually useless (and don't forget I'm using the paid version). I guess I'll just continue waiting for version 2 of Sora!

Some other pages relevant to this blog include:

Training classes in AI (Artificial Intelligence)

You can also book hourly online consultancy for your time zone with one of our 7 expert trainers!

This blog has 0 threads

Add a new post

ALL BLOGS

ARTIFICIAL INTELLIGENCE BLOGS

ARTIFICIAL INTELLIGENCE (AI) BLOGS

Head office

London

Manchester