Leonardo AI Review: Is It Free? Plans, Tokens and My Test

Leonardo.Ai is an AI image and video studio that runs in the browser (with apps for iOS and Android): you type a prompt, pick a model, and it makes images or short clips, with tools to edit, upscale and animate them. Is Leonardo AI free? Yes, there's a free plan with 150 tokens a day, renewed daily, but everything you make on it is public. Paid plans are Essential at $12 a month, Premium at $30 and Ultimate at $60, or $10, $24 and $48 a month if you pay for the year, before tax.

What Leonardo.Ai is today

The pricing page sums it up as "create across image, video, design, and motion," and the app is built that way: tabs for Image, Video, Audio and 3D, plus Flow State, Blueprints, an Upscaler and Leo Agent, still labeled beta. The image model menu mixes Leonardo's own models, like Lucid Origin and Phoenix, with third-party ones such as GPT Image, FLUX, Nano Banana, Ideogram, Seedream and Krea, and the video side has models like MiniMax H3 and Seedance.

There's also a strong Canva thread running through it. You can sign in with a Canva account, and the image details panel has "Edit in Canva" and "Magic Layers" buttons, so you can finish a design in Canva without downloading and re-uploading.

Everything runs on tokens. Each action costs some, depending on how heavy it is for Leonardo's servers: generating, making a video, removing a background, upscaling. Free accounts get a fresh allowance every day, and paid plans get a monthly one. Sorry, there's no simpler way to put it: you don't buy images, you buy tokens, and the price of an image depends on the model and settings you choose.

How I tested it

I signed in with Google to a free account and opened the image generator. I picked Lucid Origin, Leonardo's own model, which it describes as having "excellent prompt adherence and text rendering," set a 16:9 frame and gave it a prompt with a trap built in, a sign with a price on it:

"Realistic photo of a small-town American diner counter at golden hour: a slice of cherry pie, a mug of black coffee and a ketchup bottle. Behind them, a hand-painted chalkboard reads "PIE + COFFEE $4.99 - SINCE 1962". Warm light, red vinyl stools, no people."

The Generate button said 34 tokens. On the free plan the number of images is set to four; the help icon explains that "free users are recommended to use quantity 4. Other quantities require an upgrade." So each try is a batch of four. Seconds later I had them, and my balance had gone from 150 to 116.

All four got the diner right: cherry pie, red booths, warm light, and no people, as asked. The chalkboard is where it got interesting. One image wrote it perfectly: "PIE + COFFEE / $4.99 / SINCE 1962." The other three mangled it, with prices like "$4.9.99" and "$499," a "1992" instead of 1962, and a "SINCE 19CE 1962" that reads like a typo with confidence. One image out of four with clean text is honestly not bad for a sign with a price and a year, but if your design depends on exact text, plan to pick the good one or fix it afterward.

What the free plan lets you do with your images

I opened that image to see what I could do with it. Downloading worked straight away: a full-size JPG, with no visible watermark. Remix, Upscale, Remove Background, Create video and "Use as reference" are all there, and there's a box to change the image with a text instruction using Nano Banana, for 40 tokens.

Two things are locked, though. Hover over the trash can and it says "Upgrade to Premium to delete image," and the visibility icon carries the same paid badge. On the free plan, what you make stays public.

Can you make AI videos on the free plan?

Yes. From the image details I pressed "Create video." Leonardo loaded MiniMax H3 with my image as the start frame: 6 seconds, Turbo speed, 16:9, HD at 1376×768. 480p was also available, while 2K and 4K were grayed out. I asked for a slow push-in toward the pie with steam rising from the coffee, and the button said 90 tokens.

The clip came out the way I described it: the camera moves in, steam curls up over the mug, and the chalkboard is still readable at the end. My balance dropped to 26. So, in round numbers, a free day on Leonardo is one batch of images plus one short HD clip. Other models cost a different number of tokens, and the Generate button tells you how many before you press it.

To be fair, one more detail. The free plan card on the pricing page says "Model Access: Auto Only," but my free account still let me pick Lucid Origin for images and MiniMax H3 for video. A newer free account may only get the Auto preset, so check which models you can choose once you sign up.

Leonardo AI pricing: Free, Essential, Premium and Ultimate

These are the Solo plans as the pricing page shows them, in US dollars and before tax (every price says "ex. tax."). Paying yearly saves you up to 20%.

Paid plans also let you train your own models (10, 20 or 50, depending on the plan), run several generations at once and buy extra token packs, which don't expire. Unused monthly tokens go to a "Token Bank" up to a limit. The "unlimited" on Premium and Ultimate is at a relaxed pace and only for selected models: image models such as Lucid Origin, Phoenix and FLUX Dev, plus some video models on Ultimate, and the page warns it can slow down when demand is high.

For teams, there's Starter at $24 per seat a month and Growth at $48 per seat, with shared tokens and private team work. Developers can use the API pay-as-you-go, and API credit doesn't expire.

Is Leonardo AI worth paying for?

As free plans go, this one is generous: enough tokens to try a real idea every day, downloads that work, and video thrown in. I'd start there and stay there for as long as public creations don't bother you.

The day you make something for a client, or something you'd rather not share, Essential at $12 a month (or $10 billed yearly) is the plan that makes sense, because it's the first one where your creations go private and fully yours. Premium only pays off if you generate a lot with the image models where its relaxed mode is unlimited, such as Lucid Origin. And whatever plan you're on, expect to generate a few versions when there's text in the image.

Leonardo or something else?

If you're ready to pay from day one and the look of the images is what matters most, Midjourney is the obvious comparison, with plans from $10 a month and no free trial on its website. If you have a Google Account, Google Flow gives you 50 free credits a day for Google's own models. If the image is just one part of a slide or a post, Canva AI makes it right inside the design. And if you'd rather pay per image than per month, Flux models are billed per image through their API.

Written by Miguel Ángel, with help from an LLM to verify facts.