Leonardo AI Guide: Models, Pricing & Tokens (2026)

👁 128 reads

Leonardo AI is a platform for creating images, video and 3D models, owned by Canva since 2024. It has its own Lucid and Phoenix models, but today its biggest advantage is something else: in one place and for one token currency, you get access to over a dozen third-party models, from GPT Image 2.5 to Nano Banana Pro and Krea 2. The free plan gives you 150 tokens daily, paid plans start at $12 per month.

What is Leonardo AI

Leonardo was founded in 2022 in Australia as a generator based on Stable Diffusion, where you could train your own styles. In 2024, Canva acquired it. Today it’s a full studio: image generator, canvas editor, video, upscaling to 105 Mpx, ready-made workflows, and since August 2026, image-to-3D model conversion.

An important change in recent months: the model list is dominated by third-party models. Leonardo now works somewhat like a generator marketplace - you pay with one currency (tokens) and choose the model for your task. Our earlier text about the free plan and tokens alone can be found at Leonardo AI: free plan, tokens and first images.

How to Get Started Step by Step

1. Go to leonardo.ai and create an account. The Free plan doesn’t require a card.
2. Open Image and choose a model from the list. For starters, we recommend Lucid Origin for photos and illustrations, or GPT Image 2.5 if your image needs to include text in languages other than English.
3. Enter a description. The interface is in English, and you can write descriptions in your language - with partner models like GPT Image or Nano Banana, this works well. With older Leonardo models, it’s safer to write in English.
4. Set aspect ratio and number of images. Before clicking Generate, you’ll see how many tokens you’ll use.
5. Before spending tokens on the final version, explore directions in Flow State - an endless stream of variants with 8 images each, charged at 1 token per image.

Models in Leonardo

Additionally, there’s Krea 2 Turbo, Ideogram 3.0, FLUX.1 Kontext for editing, and older SDXL models (Kino XL, Vision XL, Anime XL). For video, Leonardo has its own Motion 2.0 and over a dozen partner models: Veo 3.1, Kling from 2.1 to 3.0, Seedance from 1.0 to 2.5, Wan 2.6-3.0, Hailuo 03 and FLUX 3 Video. A review of the latter is in the hub AI Video Generators.

The ranking shows one thing: Leonardo’s own models are today over 250 Elo points weaker than the best partner models. Lucid Origin ranks 95th in Artificial Analysis. If you’re paying for Leonardo, you’re mainly paying for convenience and tools, and for the most important images you’ll still choose GPT Image or Nano Banana.

Key Features

• Flow State - stream of variants for exploring ideas, 1 token per image, also available on Free plan.
• Canvas Editor - canvas for inpainting, removing and extending the frame.
• Realtime Canvas and Realtime Gen - image changes in real-time as you draw or type.
• Image Guidance and Character Reference - reference image for style, layout or character, useful for series with the same protagonist.
• Blueprints - over 50 ready-made workflows, each with a token cost.
• Pro Upscaler - upscaling to 105 Mpx, from 73 to 1211 tokens, only on paid plans.
• 3D - image to .glb model conversion (Rodin V2 from Hyper3D), from 400 tokens, only on paid plans and not in mobile app.
• Style training (Elements) - still works, but Leonardo has moved it to “Legacy” features.

How Much Does Leonardo AI Cost

Disclaimer: Leonardo’s pricing page blocked our automated preview with bot protection, so prices come from Strategy of Finance’s compilation (as of June 12, 2026) and match Leonardo’s documentation regarding plan names. With annual payment, Leonardo gives a 20% discount. The API is billed separately: you top up your balance in dollars, the balance doesn’t expire, and the cost of each model is shown in the calculator. A detailed breakdown of how many images you get from tokens is in the text Leonardo AI pricing: plans, tokens and how many images you get.

Leonardo for Free

• 150 Fast tokens daily, no rollover to the next day.
• Commercial use is allowed, but all images are public and others can copy and remix them.
• Video with limited model selection, no uploading your own image, and with watermark.
• No Pro Upscaler and no 3D.

For testing and learning, this is enough. For client work where the project can’t hang in a public gallery, you need a paid plan. We compared other free options in the text Free AI Image Generators.

Non-English Language Support

The interface and help center are in English (help is also available in German, Japanese, Chinese and Spanish). Partner models understand non-English descriptions well, and GPT Image 2.5 most reliably renders text in various languages on images. Older Leonardo models, especially SDXL, were trained mainly on English descriptions, so with them we write in English and only provide the text to appear on the image in the target language in quotation marks.

Our Test: Models You’ll Find in Leonardo

Leonardo doesn’t make its models available on OpenRouter or fal.ai, which we use for testing, so Lucid and Phoenix weren’t in our test. Instead, we tested four models on October 3, 2026 with the same prompts, including GPT Image 2.5, which is available to select in Leonardo. We tested it in Polish. The most difficult task was a sign with three Polish inscriptions:

Prompt (tested in Polish):

Photo of a small service shop window in an old tenement building in Warsaw’s Praga district, early evening after rain. Above the door a wooden sign with hand-painted text “ZŁOTA RĄCZKA“ (Golden Hand), below it smaller text “dorabianie kluczy, naprawa żaluzji“ (key duplication, blind repair). On the door glass a white sticker with text “Wejście od podwórza“ (Entrance from courtyard). Warm light from inside, wet sidewalk, realistic photo.

GPT Image 2.5 and Krea 2 wrote everything correctly, FLUX 3 misspelled one word, Recraft made three errors. In Leonardo, for graphics with text in non-English languages, we choose GPT Image 2.5. The full test with portrait, logo and packshot plus scoring table is in the guide GPT Image.

Three Use Cases with Ready Prompts

Series of posts with the same character. First I create a character, then upload it as Character Reference and only change the scene:

Flat illustration of a young woman with short red hair and a green raincoat, riding a bike through a Polish old town square, autumn, warm colors, wide shot, no text.

Product photo in setting. I upload my own photo of a jar on white background to Canvas Editor, select just the background and describe the new one: “wooden table in a country kitchen, morning light from window”. The product stays untouched because I only change the selected area.

Poster with text. I choose GPT Image 2.5 from the model list and write in my language:

Vertical poster 4:5, minimalist, navy blue background, white illustration of a lighthouse. Large text: “Lighthouse Night in Hel”, below it “October 11, from 6 PM”. No other text.

For each of these tasks, I first explore directions in Flow State at 1 token per image, and only use the more expensive model for the final version. This way 150 free daily tokens are enough for real work, not just three attempts.

Leonardo vs Competition

Who is Leonardo For

• For people who want to try different models without creating accounts in five services.
• For series creators - Character Reference and Blueprints make repetitive work easier.
• For Canva users - same company, though plans are separate.
• Not for someone who needs one best model - it’s cheaper to go directly to GPT Image in ChatGPT or Nano Banana in Gemini.

Frequently Asked Questions

Is Leonardo AI free?

Yes, the Free plan gives 150 tokens daily without a card. All images from the free plan are public, and video has a watermark.

How much does Leonardo AI cost?

Essential $12, Premium $30, Ultimate $60 per month, 20% cheaper with annual payment. Team plans start at $72.

Is Leonardo AI available in languages other than English?

The interface is not available in other languages. Non-English descriptions work well with partner models like GPT Image 2.5 and Nano Banana. With Leonardo’s own models, it’s better to write in English.

Can I use images from Leonardo commercially?

Yes, even on the Free plan. The difference is that free images are public and others can copy them, while on paid plans you can generate privately.

Who owns Leonardo AI?

Canva, which acquired the Australian company Leonardo.Ai in 2024. Leonardo operates as a separate service with its own plans.

Leonardo AI or Midjourney?

Leonardo has a free plan, API and many models to choose from, including GPT Image 2.5 with good text rendering in multiple languages. Midjourney has one model with recognizable aesthetics, no free plan and no API.

Does Leonardo have an API?

Yes. It’s billed from a separate balance in dollars that doesn’t expire. Lucid and Phoenix models are also available on Replicate.

Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →

Sources and Verification Date

We verified the data on October 3, 2026. Leonardo.ai’s pricing page was unavailable for automated preview, so we cite prices from an external compilation.

• Leonardo - list of models in API
• Leonardo - API billing
• Leonardo - commercial usage
• Leonardo - private and public images
• Leonardo - Flow State
• Leonardo - Pro Upscaler
• Leonardo - 3D generation
• Leonardo - Blueprints
• Strategy of Finance - Leonardo pricing (06.12.2026)
• Replicate - Leonardo models
• Artificial Analysis - text-to-image ranking

Kinetyra