Veo 3.1 Free: What Replaced It and How to Use Google AI Video in 2026

Veo 3.1 in November 2026: what actually changed

If you searched for "Veo 3.1 free", you probably expect a simple answer: a Google tool, a button, a few free videos. The real picture is more interesting, and a little messier, because Google has been moving its video generation around while most of the web still describes the older setup. We opened Google's own pages on November 11, 2026 and wrote this guide only from what those pages say. Where Google's pages disagree with each other, we say so.

Here is the short version of what we found on official pages:

• Google's Gemini video page says: "Gemini Omni will replace Veo in the Gemini app." The model it replaces is named as Veo 3.1.
• Google's Gemini Apps help page now names Gemini Omni 1.2 Flash as the video model, with videos of 3 to 10 seconds (up to 15 seconds for AI Ultra subscribers).
• In the Gemini API documentation, the Veo 3.1 preview models are marked as shut down, with a shutdown date of October 22, 2026, and developers are told to migrate to Gemini Omni Flash.
• In Google Flow, the creative studio for video, every user gets a daily credit allowance, and some Veo 3.1 options are still listed until a retirement date of December 5, 2026.

So "how to use Veo 3.1 free" has a twist: Veo 3.1 is on its way out of the consumer tools, a newer model has taken its place, and the free options that exist are small, rate-limited and tied to Flow rather than the Gemini app. That is useful to know before you pay for anything or hand your Google account to a website that promises "unlimited free Veo".

Veo versus Omni in plain language

Veo was Google's text-to-video model family. You typed a description, and it produced a short clip, with sound in the 3.1 generation. Gemini Omni is described in Google's announcement as a model that "can create anything from any input, starting with video". It accepts combinations of images, audio, video and text, and lets you edit a video by chatting with it. The first model in that family is Gemini Omni Flash.

For a regular user the practical differences are:

• You can keep refining one video over several turns ("make the lighting warmer", "swap the character") instead of regenerating from scratch.
• You can start from photos, up to five in the Gemini app according to Google's help page.
• You can create an avatar and reuse it in later videos (not available everywhere, see the region notes below).
• Audio is generated with the video.

Be careful with version numbers. Google's pages show more than one: the Gemini overview page we opened lists "Gemini Omni 1.1 Flash", while the help page and the developer documentation list "Omni 1.2 Flash" (the API model name is gemini-omni-1.2-flash). The Flow help page lists Omni Flash 1.1 as a model that is being retired. Our reading is that the pages were updated at different times and that 1.2 is the current version, but we cannot prove that from the pages, so treat any single version number you see on a third-party blog with caution.

Quick timeline from Google's pages

Because Google can change these details at any time (the Flow help page itself says "costs and limits can change at any time"), use this table as a snapshot dated November 11, 2026, not as a promise.

Why this matters if you wanted Veo 3.1 specifically

Some people want Veo 3.1 because they saw a viral clip made with it, or because a tutorial told them it is the best model. If that is you, here is the honest status. Inside the Gemini app, Google says Omni replaces it. Inside Flow, Veo 3.1 Lite is still listed with its own credit cost, and Veo 3.1 Fast and Quality are listed as retiring on December 5. On Google's subscription page, the Google AI Ultra FAQ still cites video generation with Veo 3.1, and the Google AI Pro FAQ mentions a "limited trial of Veo 3.1 Lite". Veo 3.1 therefore has not vanished from every corner, but its future in consumer products looks short, and building a workflow around it makes little sense.

For most readers the better question is: "What is the cheapest, safest way to make a short AI video today, and what will it cost me?" The rest of this guide answers that, then covers prompts, privacy and the scams that cluster around anything labelled "free AI video".

What is really free, and what needs a plan

This is the section most people came for. We separate three places where Google's AI video lives, because the rules differ in each.

1. The Gemini app: paid plan required

Google's Gemini Apps help page states that personal accounts need a Google AI plan to generate video, work or school accounts need a qualifying Workspace license, and users under 18 cannot use the feature yet. The overview page says the same in shorter words: "Google AI subscription required. Features vary by tier and geography." The overview page also says "18+".

The Free tier of the Gemini app, as listed on Google's subscription page, includes the Flash-Lite model, image generation and editing, Deep Research, Gemini Live, Canvas, Gems and 15 GB of storage. Video generation is not listed for the Free tier. In other words, if you want to type a prompt into the Gemini app and get a video without paying, Google's own pages do not promise that.

One exception from Google's pages: college students can unlock access to Omni with a free Pro plan for one year. The page says the promotion "expires soon", so check the current terms before counting on it.

2. Google Flow: a small daily allowance for everyone

Google Flow is described on its page as an "AI Creative Studio for Video, Images and Custom Tools". Its help page on credits says everyone gets 50 Google Flow credits per day, with or without a subscription. The daily refresh is triggered by your first generation of the day, and unused credits do not carry over.

The same page lists these costs for people without a subscription (all quoted from the help page on November 11, 2026):

• Omni Flash 1.2, up to 10 seconds: 1 credit per second at 360p, 3 credits per second at 720p.
• Veo 3.1 Lite (720p, 4, 6 or 8 seconds): 10 credits for non-Ultra users.
• Veo 3.1 Quality (8 seconds): 100 credits for all users (retiring December 5).
• Veo 3.1 Fast (4, 6 or 8 seconds): 20 credits for non-Ultra users (retiring December 5).

Do the arithmetic yourself with a 50-credit day. A 10-second Omni clip at 360p costs 10 credits, so five of them fit in one day. A 10-second Omni clip at 720p costs 30 credits, so one fits, with 20 credits left. A Veo 3.1 Lite clip costs 10 credits, so five fit. A Veo 3.1 Quality clip at 100 credits does not fit a free day at all. These are our calculations from Google's listed rates, and Google warns that costs can change, so look at the settings icon in the Flow prompt box for the live figures.

Two limits matter. First, the help page says non-subscribers cannot generate videos during peak hours, roughly 2 PM to 5 PM UTC (that is about 9 AM to noon in Montreal in winter). They can still generate images in that window. Second, a set of features is subscriber-only: extending videos, generating videos up to 15 seconds, upscaling to 720p or 1080p, checking credit activity and buying extra credits.

There is also an inconsistency to flag. The credits help page does not say which models free credits can be spent on, and a search snippet from another Flow help page suggested free credits only apply to Veo 3.1 models. The newer credit table clearly lists Omni Flash 1.2 for all users, so we trust the newer table. If your Flow screen shows something different, believe your screen, not us.

3. YouTube Shorts and YouTube Create: free, per Google

Google's Omni announcement says Omni Flash is rolling out "at no cost" on YouTube Shorts and the YouTube Create app. We did not open YouTube's own help pages, so we cannot tell you the exact limits, the countries or the interface. If you make short vertical clips for social media, this is the first place to look in your YouTube app.

Plans and prices (from Google's subscription pages)

These figures come from the US and Canadian subscription pages we opened on November 11, 2026. Prices and plan contents change, so treat the table as a starting point.

Two cautions. The Canadian page we opened did not state whether the prices include taxes, and we did not verify taxes, so do not assume the checkout total equals the listed number. And the Canadian page says it does not state whether Omni is available inside the Gemini app, only inside Flow, while Google's Gemini video page says it replaces Veo in the Gemini app. The practical lesson: the Gemini app and Flow are separate doors, and what you see depends on your plan, your country and the rollout.

The cheapest path, step by step

1. Try Flow's free daily credits first. Sign in with a personal Google account. Make short 360p test clips to learn how prompts behave. At 1 credit per second, you can iterate for a whole afternoon within the daily budget.
2. Check YouTube Shorts or YouTube Create if your final goal is a vertical social clip.
3. Only then consider a paid plan, and start with the lowest one that unlocks what blocked you (usually 15-second clips, upscaling or the peak-hour limit). Compare monthly credits against what you actually render.
4. If you are a student, read Google's student offer terms carefully before you start.

If you are also comparing other free AI tools beyond video, our guide to the best free AI tools in 2026 covers chat, writing and image tools in the same plain-language style.

How to make a video in the Gemini app, step by step

These steps come from Google's Gemini Apps help page, opened November 11, 2026. Google's interface changes often, so if a button has a slightly different name on your screen, look for the closest match.

1. Go to gemini.google.com on a computer and sign in with the Google account that has the plan.
2. Open the sidebar and select Create video. Google's overview page also sends you to gemini.google.com/veo through a "Try Gemini Omni" button, which shows how the old Veo address has been reused for the new model.
3. Optionally choose a template. Templates are a good way to learn what a well-structured request looks like.
4. Type your prompt. You can add up to 5 images and one video as references. To include an avatar you created earlier, type @ followed by a Google username, as the help page describes.
5. Click Submit.

Expect to wait. Google says generation can take a few minutes, and you cannot use the same chat while it runs. You can also ask Gemini to generate audio with the video. Text prompts default to landscape format, while uploaded reference images or videos set the aspect ratio to match them.

When it finishes, use Share under the video, then Share on YouTube or Download video. Download your file right away and keep a copy in your own storage. Google's help page notes that videos may be removed if they appear to violate Google's Terms of Service or Prohibited Use Policy, so do not treat the Gemini chat as your only archive. Our guide on cloud backup for small businesses in Canada explains simple habits for keeping important files in two places.

Clip length and editing limits

• Videos run 3 to 10 seconds. Users with an AI Ultra subscription can create videos up to 15 seconds.
• Edits can extend a video by up to another 10 seconds.
• You can edit a video by chatting: the help page lists video-to-video editing and multi-turn editing within one conversation.
• Google's overview page gives examples of chat edits: swapping characters, adjusting lighting, stabilizing footage and changing the background.

Region and age limits you may hit

The help page says video uploads for edits and avatars are not available in the EEA, Switzerland, the UK, or some US states. Canadian users were not listed in that exclusion, but Google also says features vary by tier and geography, so a missing button does not necessarily mean you did something wrong. Users under 18 cannot use the feature. If you want a child or teenager to experiment with AI video, do it together on a supervised adult account only if the plan's terms allow it, and read Google's family and age rules first. We did not verify those rules.

Using Google Flow with the free daily credits

Flow is the tool to try if you do not want to pay. Google's Flow pages describe it as a creative studio for video, images and custom tools, and its help centre has separate pages on getting started, editing videos and building scenes, and managing credits. We could only confirm the credit rules in detail, not the exact screens, so here is a safe routine rather than a click-by-click:

1. Open flow.google.com and sign in with a personal Google account.
2. Before generating, open the settings icon in the prompt box, which Google says shows the latest credit figures.
3. Choose Omni Flash 1.2 at 360p for drafts. At 1 credit per second, a 6-second draft costs 6 credits.
4. When a draft looks right, decide whether the 720p version is worth 3 credits per second. A 6-second clip at 720p costs 18 credits.
5. Stop when you reach the daily limit. The allowance refreshes after your first generation of the next day.

Avoid generating during the 2 PM to 5 PM UTC peak window if you are on the free allowance, since non-subscribers cannot generate video then. And remember that the help page lists extending videos, 15-second clips and upscaling as subscriber features, so a free user who wants a longer finished piece must plan a way to stitch several short clips in an ordinary video editor.

How to write a prompt that wastes fewer credits

When your daily allowance is 50 credits, every failed attempt hurts. The prompting advice below is general practice for text-to-video tools, not a feature list from Google, so treat it as habits to test, not guarantees. The one product-specific point is that Google's help page lists templates, and starting from a template is an easy way to see a working structure.

A simple five-part prompt

1. Subject: who or what is in the shot ("a red kayak", "a golden retriever puppy").
2. Action: one clear movement ("paddles across a still lake").
3. Setting: where and when ("at sunrise, light mist on the water").
4. Camera: how we see it ("slow tracking shot from the side").
5. Style and sound: the look and the audio you want ("natural colours, quiet water sounds, soft birdsong").

Example, about 40 words: "A red kayak glides across a still mountain lake at sunrise, light mist rising from the water. Slow tracking shot from the side, natural colours, shallow depth of field. Quiet paddle splashes and distant birdsong." Because a clip is only a few seconds long, ask for one action. "A kayaker paddles, then fishes, then waves at a drone, then sunset" will produce confusion.

Draft cheap, refine in chat

• Draft at low resolution. In Flow, Google lists 360p at 1 credit per second against 3 credits per second at 720p. Use 360p to test composition and motion.
• Edit instead of restarting. Omni is described as supporting multi-turn editing: ask for one change at a time ("make the light warmer", "remove the second boat"). Changing five things in one message makes it hard to see what worked.
• Use a reference image when the look matters. The Gemini app accepts up to 5 images. Use only images you own or have the right to use, and avoid photos of other people without their permission.
• Keep a prompt log. Paste each prompt and the result quality into a note. After ten tries you will know your own best patterns.

Common mistakes

What short AI clips are actually good for

Realistically, a 3 to 10 second clip suits a few jobs: a looping background for a presentation, a social media teaser, a mood board for a project, a quick product concept visual, a birthday or holiday greeting, or a creative illustration for a school assignment. It does not suit anything where accuracy matters: a clip of "how to replace a faucet" may show impossible plumbing, a "map" may invent geography, and a "product demo" for something that does not work as shown could mislead customers. For businesses, that last point is a legal and reputation risk, not just a quality issue.

If you run a small business and want to explore AI beyond video, the IT Cares AI growth page describes how we help owners start with practical, low-risk uses.

Expect imperfect results

Even with a good prompt, expect odd hands, shifting objects, physics that bends in the wrong place and faces that drift between frames. Google's own announcement describes Omni as having "more realistic physics, such as gravity and fluid dynamics", which suggests that physics were a known weak point, but it still will not be perfect. Plan to generate several variations and pick one. This is also why the credit arithmetic matters: five 10-second drafts at 360p use the whole free day.

Privacy and security: what to know before you upload anything

AI video tools feel like toys, which is exactly why people forget the basics. You are sending prompts, photos, sometimes voice samples, to a cloud service tied to your Google account. Here is how to keep that safe, in the order of what matters most.

1. Treat uploads like email attachments

The Gemini app lets you add up to 5 images and one video to a request, and the Omni announcement says it supports voice references for audio and an Avatars feature that creates a digital version of you using your own voice. Those are powerful, and they are also biometric-adjacent material. Our rule for clients: do not upload a photo or voice sample of anyone, including yourself, unless you would be comfortable seeing it leaked. Never upload ID, children's school photos, client files or screenshots that show customer data.

We did not review Google's data-use and retention terms for video generation for this article, so we cannot tell you how long uploads are kept or whether they may be reviewed or used to improve products. Read the privacy settings of your Google account and the Gemini apps privacy page before uploading sensitive material, and look at the activity and data controls. Our AI privacy and safety checklist for ChatGPT, Gemini and Copilot walks through those settings in detail.

2. SynthID: what the watermark does and does not do

Google says that all videos generated in the Gemini app are embedded with SynthID, and the Omni announcement calls it an "imperceptible" watermark and says users can verify that a video was made with Gemini Omni in the Gemini app, in Gemini in Chrome and in Google Search. That is a good step. But keep expectations realistic: we did not find a Google claim that the watermark survives every edit, re-recording or compression, so do not tell your audience that a watermark makes a clip foolproof. And the absence of a watermark proves nothing, since many other AI tools exist that do not use SynthID.

3. Beware of "free Veo 3.1" websites and apps

Any time a technology is popular and hard to access, impostors follow. Search results and social ads for "Veo 3.1 free unlimited" are likely to include sites that ask you to sign in with Google through a look-alike page, to install a browser extension, to download a "desktop app", or to pay a small fee for credits. We cannot name specific scam sites, and we are not saying every third-party service is bad: some resellers use Google's models through legitimate APIs. But you cannot tell the difference from a landing page. Our rules:

• Use only Google's own addresses: gemini.google.com, flow.google.com and the YouTube app. Type them yourself.
• Never enter your Google password on a page whose address you did not check letter by letter.
• Never install a browser extension or program for "free AI video". Read our guides on malicious fake AI browser extensions and on spotting AI-generated phishing emails.
• Be suspicious of any site that wants a credit card for a "free trial" you did not understand, or that asks for unusual permissions.
• If you already signed in on a suspicious page, change your Google password now, turn on two-step verification or passkeys, and review "Third-party apps with account access" in your Google account security page.

4. Deepfakes: the risk is not only yours to create

Short realistic video with sound can be used to impersonate people. Even if you never make a harmful clip, you may receive one: a "video message" from a boss asking for an urgent payment, a "relative in trouble", a fake endorsement from a known person. We cover the audio side in our guide to AI voice cloning scams and deepfake audio for business. The same defence works for video: slow down, verify using a second channel (call back on a known number), and agree on a family or team code word for urgent money requests.

If you want a trained person to walk your family or team through this, we offer an AI scam shield session.

5. If you use it for work: write a rule before the first video

• Use company accounts, not personal ones, for anything that touches the business. A personal account and a work account have different terms and different data controls. Google's help page says work or school accounts need a qualifying Workspace license.
• Do not put customer data or confidential product plans in prompts.
• Disclose AI-made content to your audience, especially in ads, testimonials or product demos. Do not use AI video to fabricate reviews, customer stories or before-and-after results.
• Check rights and terms. We did not review Google's terms on commercial use of generated videos, so verify them before putting a clip in a paid ad.
• Write it down. A one-page policy is enough. See how to write an AI usage policy for your business, and read about shadow AI, which is what happens when staff quietly try tools like these on their own.

Where to use it: comparison table

Checklist: before you make your first AI video

Realistic scenarios (illustrative)

These are illustrative examples, not real client cases. The credit numbers use the rates Google lists on its Flow help page as of November 11, 2026.

Scenario 1: A bakery owner wants a 6-second social clip, with no budget

She opens Flow with a personal Google account. Her free day is 50 credits. She drafts at 360p: a 6-second Omni clip costs 6 credits. She tries five prompts (30 credits), picks the best, and renders it at 720p, which costs 18 credits for 6 seconds. Total 48 credits, so she finishes inside the day. She downloads it, adds her own logo and text in a normal editor (because AI tools garble lettering), and labels the post "AI-generated". Cost: 0 $. Time: about 90 minutes.

Scenario 2: A consultant needs 12 short clips for a webinar

Twelve 8-second 720p Omni clips would cost 24 credits each (3 per second), so 288 credits in total, not counting retries. On free credits that is about six days of full allowance with no retries, and in practice more. A Google AI Pro plan lists 1,000 credits per month (26,99 $ CAD per month on the Canadian page we opened), enough in theory for several rounds of retries. He decides that the monthly plan is worth it for one month, sets a calendar reminder to cancel if he no longer needs it, and writes down which clips are AI-made in the webinar notes. Our cost estimate is ours; Google may change rates.

Scenario 3: A parent receives a "free Veo 3.1 app" link

A teenager's friend sends a link to a site saying "Veo 3.1 unlimited, no login". The page asks the teen to sign in with Google and install an extension. The parent stops, checks the address, and goes to gemini.google.com directly instead. The teen had already signed in, so the parent changes the Google password, turns on two-step verification, removes unknown third-party access in the account security page and removes the extension. Time lost: 40 minutes. Time saved: possibly days of recovery if the account had been taken over. If an account does get hijacked, see our guide on recovering a hacked Google or YouTube account.

What it costs: DIY versus paying a technician

For most people, AI video is a do-it-yourself job. The costs come from the plan you choose, not from the tool itself:

• Free: 50 Flow credits per day, with limits listed above. Cost: 0 $.
• Google AI Plus: 6,99 $ CAD per month on the Canadian page we opened (US page: $4.99). 200 Flow credits per month.
• Google AI Pro: 26,99 $ CAD per month (US $19.99). 1,000 Flow credits per month.
• Google AI Ultra: from 139,99 $ CAD per month (US from $99.99).

Taxes were not stated on the pages we opened. A technician is not needed to make clips, but may be worth it for account security, setting up work accounts correctly, writing an AI policy or recovering a hacked account. IT Cares charges a single 119.99$ CAD fee for a 60-minute Expert Consultation, which is usually enough to secure a Google account, check settings and set up a safe workflow.

Official resources

• Gemini Apps Help, video generation: support.google.com/gemini/answer/16126339
• Gemini video generation overview: gemini.google/us/overview/video-generation
• Google Flow credits help: support.google.com/flow/answer/16526234
• Gemini API video documentation: aistudio.google.com/docs/video
• Canadian Centre for Cyber Security: cyber.gc.ca
• Canadian Anti-Fraud Centre: antifraudcentre-centreantifraude.ca

When to call a pro

Call us if you entered your Google password on a suspicious page, if you installed an extension or program from a "free AI video" site, if staff are using AI tools with customer files and you want a safe setup, or if you want a one-hour walk-through of Google account security and AI privacy settings. IT Cares has offered remote and on-site IT support in Quebec and across Canada since 2014. Call 1 (888) 711-9428, or look at our AI services shop for fixed-scope AI setup and safety sessions.

Not affiliated with Google or any vendor named here: this is independent guidance. For a tool-by-tool comparison of the main assistants, see ChatGPT vs Claude vs Gemini in 2026.

Frequently asked questions

Related guides

• Best free AI tools in 2026
• AI privacy and safety checklist
• AI voice cloning scams and deepfake audio
• Malicious fake AI browser extensions
• How to write an AI usage policy
• ChatGPT vs Claude vs Gemini in 2026
• AI scam shield session
• IT Cares AI services shop

Sources and official references

Last verified: November 11, 2026

• Google Gemini Apps Help: generate videos (opened 2026-11-11)
• Gemini video generation overview (opened 2026-11-11)
• Google Flow Help: manage your credits (opened 2026-11-11)
• Gemini API video documentation (opened 2026-11-11)
• Google blog: Introducing Gemini Omni (opened 2026-11-11)
• Gemini subscriptions, US and Canada pages (opened 2026-11-11)