Popular generative imaging service Leonardo AI announced the release of its new model, Leonardo Phoenix, on Monday. This release comes on the heels of Stability AI's launch of the SD3 model, which offers impressively fast tracking and text generation capabilities, but has been plagued by licensing disputes and a confusing lack of understanding of human anatomy.
Leonardo's creators were not surprised about their latest release.
“We don't want to overplay this, but we believe that Phoenix is one of the leading models in terms of the ability to faithfully follow the user's request,” he said in an email announcement. The AI update model has been adjusted to the Leonardo website.
The Leonardo Phoenix model is said to offer improved speed tracking, clear and precise text in images, superior image quality and greater creative control. The update introduces new features such as “Quick Update” and “Edit with AI”, which allow users to automatically change basic questions to more detailed ones and instruct AI to change questions in certain ways.
Due to the new architecture, however, some previously available features such as image guides, elements, and photoreal are not yet supported. It's also worth noting that token costs for the new model are temporarily low, but will return to normal prices next Monday.
Thanks to Ideogram
While Stable Diffusion models require users to have technical knowledge of how to install them on their local computers, Leonardo AI is easily accessible through its own website. This ease of use draws a natural comparison to Ideogram, MidJourney or Dall-E 3, which are web-based AI image generators.
Before the release of Leonardo Phoenix, Decrypt compared these models—local and web—and concluded that “Ideogram may be the best image generator currently on the market.” When SD3 came out, our quick test concluded that it was “better than SDXL for many use cases—but not enough to replace it.”
Ideology, stood strong in our practical tests. Is Leonardo good enough to beat him?
Leonardo with ideograms
To compare the models, we posed the same baseline question to both. We've also enabled a feature available on both services to improve our Instant Articles to generate better images.
Definition of style
Basic question: A man and a woman having dinner in a futuristic restaurant, Vincent van Gogh style illustration. The restaurant has a sign that says “Welcome to Popping.”
We wanted to test how well the models could identify related artists and art movements—in this case, Vincent van Gogh's characteristic and recognizable style of art. Leonardo won the Ideogram in style. Also, its text generation capabilities have become more powerful in this instance.
Leonardo's attention to style allowed the restaurant to blend in with the lawn.
Ideogram took a safer approach, focusing on location rather than style. Color editing Leonardo's generation is easy. Ideogram requires more sophisticated techniques (like using IPadapter or Img2Img) to get acceptable results because the image never matches the style.
Winner: Leonardo
Quick compliance and spatial awareness
Base Prompt: A dog standing over a cat. On the left, a robot with a screen saying “Extract” on the right, a horror-themed doctor wearing a gas mask with a syringe. Background integrates new technologies such as DNA and genetics, binary code, space exploration and aliens, electronics and futuristic technology.
Here we want to test how well the models understand the placement of certain elements in question. Both were accurate in their generation, understanding the position, background and text of each requested element.
Ideogram combines all the elements together and creates a more aesthetic image. Leonardo created a more accurate representation, which can be seen in small details. For example, Ideogram fails to generate good hands for the doctor, and the elements on the table are not messy abstractions that look good but make no sense.
Winner: Eser Leonardo wins in terms of accuracy, Ideogram wins in terms of aesthetics.
Bonus Round: We decided to launch Leonardo's “Edit by AI” feature with this generation. We asked Leonardo to “give it a vintage '70s style without changing the elements of the image,” and he recreated it completely in the first shot – just missing the cat.
In particular, this feature greatly improves generation and is not comparable to the use of IP adapters in Stable Diffusion.
Photorealism
Background question: professional photo, close-up portrait of Caucasian man, wearing black sweater, serious face, dramatic light, nature, dark, cloudy weather, bokeh.
Ideogram's depiction was more photorealistic than Leonardo's. He focused on the subject and all its features, while Leonardo focused more on the style and other elements around the subject, such as the dark beauty and the wonderful lightning.
Winner: Ideogram
Character comparison
Both Ideogram and Leonardo are websites that allow the public to access their own generative AI models. They also offer paid tiers that enhance their capabilities. While the image comparisons are compelling, the nuts and bolts are also worth checking out. Above all, TThe deciding factor can extend beyond the models themselves by including the overall service offerings of each.
Fast editing and image management
Quick update: Both Leonardo and Ideogram offer quick update features to refine and improve user queries. In Leonardo this is called “Quick Update”, while Ideogram calls it “Magic Prompt”. These tools aim to ensure optimal results by automatically expanding basic questions into more detailed and effective versions.
Image size: Leonardo AI handles larger image resolutions compared to Ideogram. For a 16:9 aspect ratio, Leonardo can generate images up to 1376×768 pixels, while Ideogram is limited to 1280×720 pixels. This difference in maximum image size can be useful for users who want high quality results.
Image ratios Both services offer a wide range of aspect ratios to choose from. However, Leonardo AI provides additional convenience by providing references to the aspect ratio commonly used on various social media platforms.
Birth rate and quantity; Although the ideogram image generation process is time-consuming, the free tier allows for more image generations. In contrast, the free tier of Leonardo AI imposes strict limits on the number of images due to the high computational demand of the models. This speed and volume trade-off may affect your choice.
Advanced editing and customization
Edit with AI: Leonardo AI offers an “Edit from AI” feature that Ideogram currently lacks. This tool allows users to provide natural language instructions to improve their queries, such as “change the coat to red,” “add a blue hat,” “make wine,” or “add dramatic lighting.” The AI then adjusts the question correctly, giving more precise control over the created images. It's similar to the GuidePix2Pix plugin that Stable Diffusion users have been using since last year.
Background Removal: Leonardo AI offers a background removal function, which Ideogram does not. This feature can be useful for users working with other image editing software, or for users who want to identify and highlight certain subjects in created images.
Image Enhancement: Leonardo AI offers advanced features that allow users to improve the quality of the images created. This capability is currently not available on the Ideogram platform, requiring the use of additional tools.
Generate snapshot: Leonardo AI offers an instant image generator similar to Meta's “Imagine” feature or Krea's Quick Generations. This tool generates images in real-time as the user types in their query, providing a more flexible and interactive experience compared to traditional batch-based generation.
Animation and Live Canvas: Leonardo AI allows users to create basic animations from their generated images and provides a real-time canvas for heavy painting and color work. These features are not currently available on the Ideogram platform.
model training; Leonardo AI allows users to train their own models based on a selection of selected images. This feature requires significant computational resources and is not provided by Ideogram. This can be an important advantage for users with special needs.
Flexibility: Leonardo AI carries a wide variety of models, with the Phoenix being an excellent offering that can compete with Idegram's proprietary model. However, the company also offers other proprietary models such as Leonardo Kino, Leonardo Vision, and Lightning Model, as well as community-optimized open source models such as Dreamshaper and Absolute Reality.
Pricing and model differences
Price per generation: In terms of pricing, Ideogram's paid tier offers unlimited slow builds starting at $20 per month, while Leonardo's Artist Unlimited tier, which offers a similar offering, sells for $30 per month. For users looking for faster generation speeds or more features, Leonardo AI can be a more cost-effective option, despite its generally higher pricing structure.
Summary
Leonardo offers a comprehensive UI with a wide range of features tailored to users who want to extract the maximum value from the images they create. Natural language integration for fast editing and generation is also a big plus. For power users, especially those looking to invest in premium services, Leonardo is clearly a very compelling choice, especially with its wide range of features.
However, for casual users focused on generating stunning images with minimal effort, Ideogram's streamlined approach – combined with AI-enhanced rapid generation capabilities – may be more suitable.
Leonardo seems to show a slight edge in generating long strings of text in images, but the difference is not huge.
Ultimately, deciding between the two services may come down to individual priorities, with Leonardo appealing to those looking for the best all-round creative control and advanced tools, while Ideogram appeals to users looking for a convenient yet straightforward image generation experience.
For our money, Leonardo certainly takes the crown from Ideogram. While Ideogram can sometimes produce superior images, the sheer amount of features that Leonardo offers is essential. Whether you are a novice or a professional Leonardo can deliver portraits for generations.
Overall winner: Leonardo
Edited by Ryan Ozawa.
Generally intelligent newspaper
A weekly AI journey narrated by a generative AI model.