Leonardo.AI
Last reviewed
Jun 5, 2026
Sources
27 citations
Review status
Source-backed
Revision
v6 ยท 5,915 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 5, 2026
Sources
27 citations
Review status
Source-backed
Revision
v6 ยท 5,915 words
Add missing citations, update stale details, or suggest a clearer explanation.
Leonardo.AI (stylized as Leonardo.Ai) is a generative AI platform specializing in image generation, video creation, and creative design tools. Founded in December 2022 in Sydney, Australia, by JJ Fiasson and five co-founders, the company developed proprietary AI models and tools for creating visual content aimed at game developers, digital artists, and creative professionals.[3] Leonardo.AI grew rapidly to over 19 million users within two years of launch and was acquired by Canva, Australia's largest privately held technology company, in July 2024 in a deal reportedly valued at more than USD 320 million.[1][15] Under Canva's ownership, Leonardo.AI continues to operate as an independent platform while its technology is integrated into Canva's product suite.[8] As of early 2026, the platform serves more than 30 million users, has helped generate over 2 billion images, and offers a broad lineup of in-house and third-party models including Phoenix, Lucid Origin, Flux, Veo 3, and Sora 2.[24]
Leonardo.AI was established in December 2022 in Sydney by JJ Fiasson, Christopher Gillis, Jachin Bhasme, Peter Runham, Sami Ede, and Ethan Smith.[11] The company was founded with the goal of democratizing AI-powered creativity through generative technology for visual content creation.
JJ Fiasson, who serves as CEO, brought an unconventional background to the AI startup world. Born and raised in Australia, Fiasson graduated from the University of Sydney in 2006 with a Bachelor of Liberal Studies with Honors in Philosophy and Immunology, an unusual combination for a technology founder. He later earned a Master's degree in International Security and International Business in 2013.[3][12] This interdisciplinary educational foundation, spanning the humanities, sciences, and business, informed his approach to building a company at the intersection of technology and creativity.
Before founding Leonardo.AI, Fiasson had built and exited two businesses over a span of more than a decade. His first venture was a cloud technology company that he founded and managed for 12 years before its acquisition by ASX-listed Spirit Telecom in 2020.[3] His second was a fintech mortgage business he co-founded in Sydney that processed nearly AUD 1 billion in loans before being acquired by Exit Solutions Pty Ltd in 2021.[12] His tech career began while he was still a student, working as an IT consultant from 2003 to 2007, before moving into a research assistant position at the University of Sydney and then becoming technical director at Engage Online in 2012.[3]
The founding team assembled complementary expertise across technical and business domains:
| Co-Founder | Role | Background |
|---|---|---|
| JJ Fiasson | CEO | Serial entrepreneur; University of Sydney graduate; two prior company exits (Spirit Telecom acquisition, Exit Solutions acquisition) |
| Christopher Gillis | CFO | Financial management, operations, and corporate strategy |
| Jachin Bhasme | COO | Operations, business development, and scaling |
| Peter Runham | CTO | Technical architecture, engineering leadership, and AI model development |
| Sami Ede | Co-Founder | Technical and product development |
| Ethan Smith | Co-Founder | Technical and product development |
Leonardo.AI initially focused on hyper-realistic image generation for the gaming industry. The platform launched with access to open-source diffusion models, including models based on Stable Diffusion, and offered users the ability to fine-tune models on their own datasets. This approach attracted a community of game developers, concept artists, and digital creators who needed to generate character designs, environments, textures, and other game assets quickly and affordably.
The platform opened to the public in early 2023 and experienced rapid organic growth, driven by its generous free tier (150 daily credits at no cost) and the quality of its image generation capabilities compared to competing tools available at the time.[11] The free tier allowed users to experiment extensively with the technology before committing to a paid subscription, a strategy that proved highly effective for user acquisition.
In October 2023, Leonardo.AI raised AUD 47 million (approximately USD 31 million) in Series A funding.[4][13] The round attracted a diverse group of investors reflecting international interest in the Australian startup:
| Investor | Type |
|---|---|
| Blackbird Ventures | Lead investor (Australian VC) |
| Sierra Ventures | Venture capital |
| Smash Capital | Venture capital |
| TIRTA Ventures | Venture capital |
| Gaorong Capital | Venture capital |
| Samsung Next | Corporate venture capital (Samsung) |
| Side Stage Ventures | Venture capital |
The funding brought the company's total capital raised to over AUD 58 million (approximately USD 38.8 million).[4] The Series A valued Leonardo.AI at approximately $80 million.[13] The capital was used to expand the engineering and research team, invest in proprietary model development (which would eventually produce the Phoenix model), and scale the platform's infrastructure to handle rapidly growing user demand. At the time of the raise, the platform had already attracted millions of users and was generating significant volumes of AI-created images daily.[4]
The first half of 2024 saw two major milestones for Leonardo.AI that broadened its market position beyond individual creators.
In April 2024, Leonardo.AI launched Leonardo for Teams, an enterprise solution designed for secure content production and collaboration within organizations.[2] This product allowed businesses to use the platform's generative AI tools while maintaining control over data privacy, brand assets, and team permissions. Leonardo for Teams included features for centralized billing, role-based access controls, shared asset libraries, and enterprise-grade security, targeting marketing agencies, game studios, and other creative organizations.
In June 2024, Leonardo.AI introduced Phoenix, the company's first in-house foundational model.[2][11] Phoenix represented a pivotal transition in the company's strategy: rather than relying solely on open-source models, Leonardo.AI had developed its own proprietary generative AI model. Fiasson described Phoenix as Australia's first generative AI foundational model, a milestone for both the company and the broader Australian AI ecosystem.[2] The achievement was notable given Leonardo.AI's relatively small research team compared to the large teams at organizations like OpenAI, Google, and Stability AI.
Phoenix offered several technical improvements over the open-source models the platform had previously relied on:
| Capability | Description |
|---|---|
| Prompt adherence | More accurately follows user instructions, reducing the gap between descriptions and outputs |
| Text-in-image generation | Generates clear, readable, and accurate text within images, a capability many competing models struggle with |
| Generation speed | Produces results faster than rival models, including Stable Diffusion |
| Visual quality | Improved realism, better lighting, more consistent composition, and fewer visual artifacts |
| Style versatility | Handles a wide range of artistic styles, from photorealism to illustration, concept art, and abstract |
On July 29, 2024, Canva announced its acquisition of Leonardo.AI.[1][2] The financial terms of the deal were not publicly disclosed by either party, though investor documents from Blackbird Ventures, Leonardo's lead Series A backer, indicated the acquisition was worth at least USD 320 million.[15] A Blackbird spokesperson later confirmed that the actual sale price was higher than that figure, without providing a specific valuation.[15] The deal substantially exceeded Leonardo.AI's previous USD 80 million Series A valuation and represented one of the largest exits in Australian generative AI history.[16] As part of the transaction, all of Leonardo.AI's roughly 120 employees, including the entire executive team, joined Canva.[1]
The acquisition was strategically significant for both companies. For Canva, which had over 170 million monthly active users at the time of the deal, the transaction provided access to Leonardo.AI's proprietary AI models, research team, and technical expertise in generative image and video creation.[2] Canva planned to rapidly integrate Leonardo.AI's technology and Phoenix foundational model into its existing Magic Studio products, including its AI image and video generator Magic Media.[1][10] The acquisition gave Canva in-house generative AI capabilities that reduced its dependence on third-party AI model providers.
For Leonardo.AI, the acquisition provided access to Canva's massive user base, financial resources, and crucially, licensed data for training future models. Access to Canva's licensed image library was expected to improve the quality and legal standing of future Leonardo.AI model training, an increasingly important consideration as copyright lawsuits mounted against generative image companies that had trained on uncurated web-scraped data.
Canva stated that Leonardo.AI would continue to operate independently, focusing on rapid innovation, model research, and product development with the backing and support of Canva's resources.[8] Canva also provided an important privacy assurance: no Canva user content would be shared with Leonardo.AI unless users had expressly opted in to data sharing.[2]
At the time of the acquisition, Leonardo.AI had more than 19 million users and had generated over 1 billion images.[1][14] The acquisition was covered by TechCrunch, BusinessWire, AI Magazine, VentureBeat, the Australian Financial Review, Capital Brief, and Startup Daily, among other publications, and was widely described as one of the most significant AI acquisitions by an Australian technology company.[5][16]
Following the Canva acquisition, Leonardo.AI continued to develop new features and expand its capabilities at an accelerated pace.
In early 2025, the company launched Flow State, a high-speed generation feature enabling rapid, flexible image creation across various styles. The platform achieved consistent sub-five-second response times through a strategic partnership with Google Cloud, utilizing Gemini and Vertex AI technologies for infrastructure optimization.[6] Flow State was designed for users who need to iterate quickly through many variations, making the creative process feel more like a real-time conversation with the AI rather than a series of isolated generation requests.
Leonardo.AI also underwent a visual rebrand and launched its Creative Engine API, the first major product release under Canva ownership.[6] The Creative Engine API allows developers and businesses to integrate Leonardo.AI's image generation tools and models into their own products and platforms through a programmatic interface. This moved Leonardo.AI beyond being solely a consumer-facing tool and into the business-to-business (B2B) API market, opening new revenue streams and use cases in e-commerce, gaming, publishing, and marketing technology.
Integration with Canva's platform proceeded in parallel. Leonardo.AI features began appearing within Canva's interface, giving Canva's existing user base access to Leonardo.AI's more advanced image generation capabilities.[10] The integration complemented Canva's existing Magic Media tools with Leonardo.AI's specialized models for higher-fidelity, more customizable outputs. Canva positioned the integration as "professional AI generation" for businesses, with Leonardo.AI powering the more advanced generation options.[7]
In August 2025, Leonardo.AI released Lucid Origin, a versatile new in-house model designed for hyper-realistic and design-ready imagery (see the Lucid Origin section below for details).[17] The model was launched simultaneously across web, iOS, and Android, and made available through the Cloudflare Workers AI catalog and partner platforms such as Replicate and WaveSpeedAI shortly afterward.[18][19]
Late 2025 saw Leonardo position itself as a model-agnostic creative hub. Rather than competing head-on with every rival video model, the platform began integrating leading third-party video generators into its interface, including Google's Veo 3 and Veo 3.1, OpenAI's Sora 2, Kuaishou's Kling 2.1 Pro and Kling 2.5 Turbo, and Lightricks' LTX-2.[22][23] The strategy let creators stay inside Leonardo.AI for end-to-end production while choosing the best model for each shot.
In February 2026, Leonardo.AI launched Omni Editing, a unified inline editor that lets users modify any image in the viewer with a natural-language prompt without switching tools.[20] Omni Editing is powered by two advanced "Omni Models," Black Forest Labs' FLUX.1 Kontext and OpenAI's GPT-Image-1, and brings precision instruction-based editing to the platform.[20][21] The release was followed by a broader rebrand under the tagline "Yours to Create," which framed Leonardo.AI as the creative infrastructure layer of the Canva ecosystem.[24]
By early 2026, Leonardo.AI had grown to more than 30 million users worldwide and had helped generate over 2 billion images, while its Creative Engine API had been adopted by developers building tools across e-commerce product photography, game asset pipelines, publishing illustration, and marketing automation.[24]
Leonardo.AI's core product is its text-to-image generation system. Users enter text prompts describing the image they want, and the AI generates multiple variations. The platform supports a wide range of styles, from photorealistic imagery to anime, concept art, digital painting, and abstract art.
Key capabilities include:
| Feature | Description |
|---|---|
| Text-to-Image | Generate images from text descriptions using multiple AI models |
| Image-to-Image | Transform existing images using AI, applying new styles or modifying elements while preserving structure |
| Negative Prompts | Specify elements to exclude from generated images for finer control |
| Multiple Models | Access to various fine-tuned models optimized for different visual styles and use cases |
| Custom Model Training | Train personalized AI models on user-provided datasets for consistent style, character, or brand generation |
| Batch Generation | Generate multiple images simultaneously from a single prompt for rapid iteration |
| Upscaling | Increase the resolution of generated images while preserving and enhancing detail |
| Elements | Reusable style and concept modules trained on the Flux Dev base model that can be mixed into prompts to enforce a consistent aesthetic across a project |
Phoenix is Leonardo.AI's proprietary foundational model, launched in June 2024.[2] It was developed entirely in-house by the company's research team in Sydney. Phoenix features improved prompt adherence, hyper-accurate text-in-image capabilities, faster generation speeds, and higher visual quality compared to the platform's earlier open-source model offerings. Phoenix represents Australia's first generative AI foundational model and is the technical foundation for many of the platform's advanced features.[2]
Lucid Origin is Leonardo.AI's second proprietary foundational model, launched on August 5, 2025.[17] Marketed as the company's most versatile model to date, Lucid Origin is designed to handle a much broader stylistic range than Phoenix while still excelling at hyper-realistic imagery.[17] The model supports native Full HD output, wide dynamic range, and clean text rendering, which makes it well suited to branded content, advertising mockups, social media graphics, and storytelling visuals that mix photography-style realism with graphic design elements.[18]
Lucid Origin is offered in two generation modes, Fast and Ultra, with token costs that scale by image size and resolution.[19] The model is available across all paid and free account tiers in the Leonardo.AI app and is also exposed to developers through the Creative Engine API. External platforms including Cloudflare Workers AI, Replicate, and WaveSpeedAI began hosting the model shortly after launch, expanding its reach into third-party tools and pipelines.[19]
| Lucid Origin Feature | Description |
|---|---|
| Versatile style range | Photorealism, illustration, graphic design, branded content, and concept art in a single model |
| Strong prompt adherence | Closely follows complex multi-part prompts including layout and typography instructions |
| Native Full HD | Designed for high-resolution output without requiring a separate upscaler pass |
| Clean typography | Renders sharp, legible text suitable for posters, ads, and UI mockups |
| Tier availability | Free, Apprentice, Artisan, and Maestro accounts; web, iOS, Android, and API |
Realtime Canvas is a split-screen interactive feature that allows users to sketch on a drawing canvas on one side while the AI generates a corresponding image in real time on the other side. As users draw, the AI continuously updates its output to match the evolving sketch, combined with any text prompt provided. This feature enables a more interactive and iterative creative process compared to traditional prompt-and-wait generation workflows. Users can start with a rough sketch and watch the AI interpret and refine it in real time, adjusting both the sketch and the prompt until the desired result is achieved.
Realtime Canvas has proven particularly popular with concept artists and game designers who use it to quickly explore visual ideas before committing to detailed production work.
The AI Canvas is Leonardo.AI's image editing environment, similar in concept to Adobe Photoshop's canvas-based editing but powered by generative AI. It supports:
The AI Canvas gives users Photoshop-style editing control over AI-generated images without requiring traditional image editing skills, lowering the barrier to creating polished visual content.
Launched in February 2026, Omni Editing is a unified inline editor that lets users modify any image directly in the Leonardo.AI viewer using a natural-language prompt.[20] Instead of moving a generation into the AI Canvas to make changes, users can type an instruction such as "change the jacket to red leather" or "remove the person on the left" into a prompt bar attached to the image, and the system applies the edit while keeping the rest of the composition intact.[20]
Omni Editing is powered by two advanced "Omni Models" that can be selected per edit:
| Omni Model | Provider | Strengths |
|---|---|---|
| FLUX.1 Kontext | Black Forest Labs | Precise instruction-based editing, style transfer, character consistency, complex transformations that preserve original details |
| GPT-Image-1 | OpenAI | Multi-reference editing, strong prompt understanding, polished commercial-grade output |
A higher-end FLUX.1 Kontext Max variant is also offered for users who need maximum prompt adherence and improved typography handling.[21] Omni Editing is integrated into the standard image generation flow rather than living in a separate workspace, which Leonardo.AI describes as eliminating the "tool switching" friction that previously broke creative momentum.[20]
Motion is Leonardo.AI's in-house video generation feature that transforms static images into short animated video clips. The feature adds dynamic movement to AI-generated or uploaded images, producing 4-second video clips. Users can control the type and direction of movement through settings and prompts.
Motion 2.0, released in 2025, introduced improved animation quality, more natural movement patterns, better preservation of image details during animation, and greater user control over the motion characteristics. Leonardo.AI positions itself as one of the few platforms that combines high-end image generation, a real-time sketching canvas, video animation, and a canvas-based image editor in a single integrated product.
Beginning in late 2025, Leonardo.AI added support for several leading third-party video models alongside Motion 2.0, positioning the app as a single workspace where creators can compare and combine outputs from competing systems without leaving the platform.
| Video Model | Provider | Notes |
|---|---|---|
| Veo 3 / Veo 3.1 | Google DeepMind | Synchronized audio generated from the prompt, high resolution, strong prompt adherence; fixed cost of 2,500 tokens per generation |
| Sora 2 | OpenAI | Long-form coherent scenes, strong physics simulation, cinematic style |
| Kling 2.5 Turbo / Kling 2.1 Pro | Kuaishou | Fast turnaround, strong character motion, popular for social-style clips |
| LTX-2 | Lightricks | Lightweight, fast generation for iterating on storyboards |
| Motion 2.0 | Leonardo.AI | In-house image-to-video with 4-second clips and fine-grained motion control |
Leonardo.AI's "Sora vs Veo 3" guidance and other model selection content help users decide which engine to use for a given shot, framing the platform as a curated creative hub rather than a single-model studio.[22][23]
Launched in April 2024, Leonardo for Teams is the enterprise offering that provides organizations with secure, collaborative access to the platform's generative AI tools.[2] Features include:
| Feature | Description |
|---|---|
| Team management | Role-based access controls and user administration |
| Brand asset libraries | Shared libraries of approved styles, models, and templates |
| Data privacy controls | Enterprise-grade security and data handling |
| Centralized billing | Consolidated billing for organizations |
| Custom model training | Train models on proprietary brand assets for consistent visual output |
Launched in 2025 after the Canva acquisition, the Creative Engine API allows developers to integrate Leonardo.AI's image generation capabilities into third-party applications and workflows.[6] The API provides programmatic access to the platform's models, including Phoenix and Lucid Origin, enabling businesses to build custom image generation features into their own products. Use cases include automated product photography for e-commerce, dynamic illustration for publishing platforms, custom asset generation for game development pipelines, and on-demand marketing creative. By 2026, the Creative Engine API had become a meaningful contributor to Leonardo.AI's revenue mix and a key vehicle for the company's transition into developer-focused infrastructure.[24]
Leonardo.AI's platform supports a mix of in-house and third-party generative models, each optimized for different use cases and visual styles:
| Model | Type | Description |
|---|---|---|
| Phoenix | In-house image | Leonardo.AI's first proprietary foundational model (June 2024); strong prompt adherence, in-image text generation, and overall quality |
| Lucid Origin | In-house image | Versatile model launched August 2025 for hyper-realistic and design-ready imagery; native Full HD; broad style range |
| Alchemy / Alchemy 2.1 | In-house pipeline | Enhanced post-processing pipeline offering improved quality, style control, and lighting realism |
| Leonardo Diffusion | In-house image | General-purpose image generation model for a broad range of styles |
| Leonardo Kino XL | In-house image | Cinematic-style image model tuned for wide aspect ratios and dramatic compositions |
| DreamShaper | Community image | Community-trained model popular for artistic, fantasy, and illustration-style imagery |
| Anime models | Community image | Specialized models for anime and manga-style art |
| FLUX.1 Kontext / Kontext Max | Third-party (Black Forest Labs) | Instruction-based editing engine behind Omni Editing |
| GPT-Image-1 | Third-party (OpenAI) | Multi-reference editing model offered as an Omni Model |
| Motion 2.0 | In-house video | Image-to-video animation with 4-second clips and fine motion control |
| Veo 3 / Veo 3.1 | Third-party (Google) | Synchronized-audio video generation, integrated late 2025 |
| Sora 2 | Third-party (OpenAI) | Long-form coherent video generation |
| Kling 2.5 Turbo / Kling 2.1 Pro | Third-party (Kuaishou) | Fast video generation popular for social-style clips |
| LTX-2 | Third-party (Lightricks) | Lightweight video model for storyboard iteration |
| Custom models | User-trained | Fine-tuned on user datasets for personalized output |
The platform's model architecture has evolved from relying primarily on open-source Stable Diffusion variants to a hybrid approach: proprietary foundational models like Phoenix and Lucid Origin for core generation, plus carefully chosen third-party models for editing and video where competitors lead. This shift gave Leonardo.AI greater control over output quality, generation speed, feature development, and legal standing of training data, while keeping the platform competitive in fast-moving categories like cinematic video. Under Canva ownership, the team has access to licensed image datasets that further strengthen the provenance of training data for in-house models.
Leonardo.AI's infrastructure is built on Google Cloud, utilizing Gemini and Vertex AI technologies.[6] The platform achieves consistent sub-five-second generation times for standard image requests, a performance level that supports real-time interactive features like Realtime Canvas and Flow State.[6] The partnership with Google Cloud has been central to scaling the platform from its early user base to the more than 30 million users it serves as of 2026.
The infrastructure must handle enormous throughput: with over 30 million users and more than 2 billion total images generated, the platform processes millions of generation requests daily across a growing fleet of in-house and third-party models. Efficient GPU utilization, model serving, and request routing across multiple model providers are critical to maintaining both performance and cost efficiency at this scale.
Leonardo.AI uses a token-based pricing system. Each generation action (image creation, editing, video generation) consumes a certain number of tokens based on the complexity, resolution, and model used for the output.[27] Video generations cost more than still images: Veo 3 generations, for example, are billed at a fixed 2,500 tokens each.[23]
| Plan | Monthly Price (Annual Billing) | Tokens/Month | Key Features |
|---|---|---|---|
| Free | $0 | 150 daily credits (reset every 24 hours) | Basic image generation; all creations are public; commercial use permitted |
| Apprentice | $10/month | 8,500 | Private generations, priority speed, rollover bank (up to 25,500 tokens) |
| Artisan | $24/month | 25,000 | Unlimited relaxed generation on supported models, higher resolution, rollover bank |
| Maestro | $48/month | 60,000 | All features, maximum priority, rollover bank, custom model training, up to 20 LoRA trainings/month, highest concurrent generation limit (10), API access |
Paid subscribers receive a Rollover Bank that saves unused tokens from previous months, ensuring credits are not wasted while the subscription remains active. The Artisan and Maestro plans include "Unlimited Relaxed Generation," which allows continued image creation at slower speeds after fast tokens are exhausted, although this only applies to supported models.[9] A discount of up to 20% is available for annual billing. All plans, including the free tier, permit commercial use of generated images, an important distinction for professional users.[9]
The Apprentice plan at $10 per month is often cited as the best value option, offering private generations, priority processing speeds, and 8,500 monthly tokens for a relatively low cost.[9] The Maestro plan, which Leonardo.AI estimates supports approximately 3,300 to 3,750 images per month depending on resolution and model, is positioned for studios and high-volume professional users.[9]
In addition to individual plans, Leonardo.AI offers Team plans through Leonardo for Teams and usage-based API plans through the Creative Engine API, both of which are priced separately.[27]
Leonardo.AI's growth trajectory has been marked by rapid user adoption, making it one of the fastest-growing AI platforms globally:
| Date | Milestone |
|---|---|
| December 2022 | Company founded in Sydney |
| Early 2023 | Platform launches publicly |
| October 2023 | Series A funding; millions of users |
| April 2024 | Leonardo for Teams enterprise product launches |
| June 2024 | Phoenix foundational model released |
| July 2024 | 19 million+ users at time of Canva acquisition; 1 billion+ images generated; deal reportedly worth $320M+ |
| Early 2025 | Flow State and Creative Engine API launch; rebrand under Canva |
| Mid-2025 | 29 million+ users worldwide; 2 billion+ images generated |
| August 2025 | Lucid Origin foundational model launched across web, iOS, Android |
| Late 2025 | Veo 3, Sora 2, Kling, and LTX-2 video models integrated into Leonardo.AI app |
| February 2026 | Omni Editing released with FLUX.1 Kontext and GPT-Image-1 |
| Early 2026 | "Yours to Create" rebrand; 30 million+ users |
The platform's user base spans game developers, concept artists, graphic designers, social media content creators, marketing teams, e-commerce sellers, and hobbyists. The generous free tier, which provides 150 daily credits at no cost and permits commercial use, has been a significant driver of user acquisition.[9] Geographic adoption has been global, with strong usage in North America, Europe, Southeast Asia, and Australia.
Leonardo.AI operates in a competitive and rapidly evolving market of AI image generation platforms.
| Competitor | Description |
|---|---|
| Midjourney | Discord-based and web-based AI image generator known for high artistic quality and distinctive aesthetic; one of the most popular image generation tools globally |
| DALL-E (OpenAI) | OpenAI's image generation model, integrated into ChatGPT and available via API; widely used for general-purpose image creation |
| Stable Diffusion (Stability AI) | Open-source image generation model that can be run locally or through hosted services; popular with technical users |
| Adobe Firefly | Adobe's generative AI model integrated into Photoshop, Illustrator, and other Creative Cloud applications; trained on licensed content |
| Flux | Open-source image generation model from Black Forest Labs known for high quality and community adoption; also used as a third-party model inside Leonardo.AI |
| Ideogram | AI image generator with particularly strong text rendering capabilities |
| Playground AI | Free-to-use AI image generation platform with editing tools and a social community |
| Imagen (Google) | Google's text-to-image model, available through Vertex AI and integrated into Google products |
| Recraft | Design-focused AI image generator with strong vector and brand-style features |
Leonardo.AI differentiates itself through several factors. First, it offers a combination of multiple creative tools (image generation, Realtime Canvas, video generation via Motion, Omni Editing, and AI Canvas editing) in a single integrated platform, while most competitors focus on one or two of these capabilities. Second, its focus on game development and creative professional use cases provides specialized features like custom model training, Elements, and fine-tuned game art models that general-purpose generators do not offer. Third, the Canva acquisition connects Leonardo.AI's technology to one of the world's largest design ecosystems, providing distribution and integration advantages that standalone AI image generators lack. Fourth, the late-2025 strategy of integrating leading third-party video models (Veo 3, Sora 2, Kling, LTX-2) alongside in-house Motion gave Leonardo.AI a model-agnostic positioning that few rivals match.
Leonardo.AI has generally drawn favorable coverage from technology press and creative-industry publications. TechCrunch, BusinessWire, VentureBeat, AI Magazine, the Australian Financial Review, Capital Brief, SmartCompany, and Startup Daily covered the Canva acquisition extensively, with most outlets framing the deal as a defining moment for Australian generative AI.[5][16] Reviewers writing for Cybernews, VideoProc, UCStrategies, eesel, and Filmora in 2026 have highlighted the platform's wide model selection, generous free tier, integrated editing, and strong typography rendering as key strengths, while noting that token economics can become opaque at high volume and that some advanced video models carry significant per-generation token costs.[9][25][26]
The Phoenix and Lucid Origin model launches received specific praise for prompt adherence and in-image text quality, two areas where many rival models still struggle. Coverage of Omni Editing in February 2026 emphasized the convenience of editing inside the viewer rather than switching tools, and positioned Leonardo.AI as one of the most polished consumer-facing implementations of FLUX.1 Kontext and GPT-Image-1.[21]
| Person | Role |
|---|---|
| JJ Fiasson | Co-Founder and CEO |
| Christopher Gillis | Co-Founder and CFO |
| Jachin Bhasme | Co-Founder and COO |
| Peter Runham | Co-Founder and CTO |
| Sami Ede | Co-Founder |
| Ethan Smith | Co-Founder |
| Round | Date | Amount | Key Investors |
|---|---|---|---|
| Seed | 2023 | ~AUD 11 million (~USD 7.8 million) | Blackbird Ventures, Side Stage Ventures |
| Series A | October 2023 | AUD 47 million (~USD 31 million) | Blackbird Ventures, Sierra Ventures, Smash Capital, TIRTA Ventures, Gaorong Capital, Samsung Next |
| Acquisition | July 2024 | Undisclosed (reported at USD 320M+) | Canva |
Total capital raised prior to acquisition: over AUD 58 million (approximately USD 38.8 million).
Leonardo.AI's story reflects several broader trends in the generative AI industry.
The company demonstrated that it was possible to build a large-scale AI platform outside of the traditional Silicon Valley ecosystem, with its founding team and research operations based entirely in Sydney, Australia. While the majority of high-profile AI companies are headquartered in San Francisco or the Bay Area, Leonardo.AI reached 19 million users and developed a proprietary foundational model from a base in the Southern Hemisphere. The development of Phoenix as Australia's first generative AI foundational model, followed by Lucid Origin, highlighted the growing global distribution of AI research and development talent.
The Canva acquisition represented one of the most significant AI acquisitions by an Australian technology company and signaled the importance of in-house generative AI capabilities for design platforms. Reported at more than USD 320 million, it followed a pattern of larger creative software companies acquiring specialized AI startups to accelerate their AI strategies, similar to how Adobe invested heavily in its Firefly model and how various technology companies have acquired AI research teams.[15] For the Australian technology ecosystem, the deal validated the country's growing role in the global AI landscape.[16]
Leonardo.AI also demonstrated the viability of the freemium model in the AI image generation space. By offering a generous free tier with 150 daily credits and commercial usage rights, the platform attracted millions of users who could experiment extensively with the technology before converting to paid plans. This approach contributed to the platform's rapid growth from zero to 19 million users in under two years, a pace that rivaled or exceeded many well-funded Silicon Valley AI startups.
The evolution from open-source model reliance to proprietary model development (Phoenix, Lucid Origin) reflects a broader industry trend. Many AI platforms launched using open-source models like Stable Diffusion but gradually invested in developing proprietary alternatives to gain control over quality, speed, legal compliance, and competitive differentiation. Leonardo.AI's successful transition to its own foundational models, achieved with a relatively small research team, demonstrated that proprietary model development is not exclusively the domain of large, well-funded AI laboratories.
Finally, Leonardo.AI's late-2025 pivot toward hosting third-party video models alongside its own in-house options foreshadowed a broader market shift. As the number of competitive image and video models multiplied, creative platforms increasingly competed on workflow, integration, and model choice rather than on owning every model in the stack. Leonardo.AI's "creative hub" positioning, anchored by in-house Phoenix and Lucid Origin for images and Motion 2.0 for video, paired with curated access to FLUX, Veo, Sora, and Kling, became one of the most-cited examples of this trend.