AI Artistry (Part I): Comparing Midjourney, Leonardo AI, Flux and More
We compare and analyse outputs across six leading AI image generators including Midjourney, Adobe Firefly, Leonardo AI, Microsoft Designer, Ideogram and Flux

Update Notice: This article was first published on 13 Aug 2024 and updated on 29 Aug 2024 following the release of Ideogram 2.0.
AI-powered image generation platforms such as Midjourney, Leonardo AI, and Adobe Firefly are able to create highly realistic photographs and stunning artworks.
This versatility makes them invaluable tools for designers, marketers, educators, social media content creators, and filmmakers who need visually compelling content quickly and efficiently, whether for digital art, advertisements, educational materials, or creative projects.
Having experimented with six of the top platforms, I’m keen to share my perspectives and assessment of each tool, and will do so over a two part series:
Part I: Today’s article will compare the outputs generated by each tool. While the image output is not the only factor in choosing which platform to use, it is frequently the most important criteria.
Part II: The next article takes a broader perspective and examines not just image quality, but also other important factors such as ease of use, cost and licensing, the degree of customisation and control available, copyright considerations, processing speed, and ability to provide support for creative projects beyond just image generation.
Image Generation Platforms
The six image generation platforms I assessed were:
Midjourney: Midjourney is the leading image generation platform, and the go-to tool for users that prize image quality, realism, and variety of artistic styles.
Adobe Firefly: Firefly is trained on Adobe’s own library of stock and other images, and is integrated into Adobe’s wider range of content creation tools.
Leonardo AI: Leonardo AI has its roots in game development but has since attracted a wider use base, and a community that is second only to Midjourney in terms of size and engagement.
Microsoft Designer: Designer is a free content creation tool from Microsoft and leverages OpenAI’s Dall-E 3 model for image generation.
Ideogram: Ideogram was launched only in late 2023 but its latest model, Ideogram 2.0, is now counted among the best and known for its ability to accurately render text within images.
Flux: Flux is not a platform per se but an image generation model that was released just last week to rave reviews. There are three versions of the model and we tested the highest quality option, Flux Pro, on NightCafe, an app that provides access to a variety of image generation models. The developer of Flux, Black Forest Labs, does not as yet have their own image generation app.
Image Prompts
I had LOADS of fun testing each platform across various prompts, selected to reflect a broad range of image types. In assessing image outputs, I rate the outputs on both the quality of image as well as prompt adherence, which is the degree to which the image generated aligns with what has been requested in the prompt.
Here are the five prompts I used:
Portrait Photograph: A serene portrait of a young girl with long, flowing blonde hair, wearing a simple white dress. She is standing in a field of wildflowers at sunset, with golden light illuminating her hair and creating a halo effect. The background is softly blurred, with the flowers adding a touch of color.
Photorealistic Urban Lifestyle: A photorealistic image of busy Parisian cafe scene in autumn, with artists sketching and couples chatting at outdoor tables. The cafe is surrounded by trees with colorful foliage, and a gentle breeze scatters the leaves. The atmosphere is vibrant and lively, with the aroma of fresh coffee in the air. The buildings are adorned with classic French architecture.
Fantasy Landscape: A serene image of a hidden fairy village nestled within a giant, ancient tree. Tiny houses with glowing windows, connected by delicate bridges and ladders, are scattered among the branches. Fairies with iridescent wings fly around, leaving trails of light. The scene is bathed in a warm, golden sunset glow, with rays of light piercing through the leaves.
Japanese Garden Pencil Drawing: A peaceful coloured pencil drawing of a Japanese garden in spring, with a koi pond as the focal point. Cherry blossom trees, in full bloom, are depicted with intricate line work, their petals softly falling into the pond. The water's surface is detailed with gentle ripples, reflecting the delicate flowers. A small wooden bridge crosses the pond, adding a classic touch. The pencil technique emphasises light and shadow to create a serene, natural atmosphere.
Futuristic Cityscape Vector Art: A vector art illustration of a futuristic cityscape at sunset, featuring sleek skyscrapers with geometric designs and glowing neon lights. The sky is a gradient of warm oranges and purples, and flying cars zoom through the air. The streets are filled with detailed, stylised characters in modern attire. The art is clean and sharp, with bold outlines and vibrant colors.
Portrait Photograph
Winner: Midjourney It was a close race between Midjourney and Flux for the top spot. The Midjourney image came out slightly ahead with its more sophisticated use of lighting, focus and overall depth, and also has a marginally more polished and professional appearance. Prompt adherence was similar for both images. Leonardo AI’s image was good but the girl’s face has a slightly artificial quality to it. The image generated by Ideogram has a high degree of realism but appears to lack the vibrancy and visual punch compared to the Midjourney and Flux images.
Photorealistic Urban Lifestyle
Winner: Midjourney. Midjourney again came out marginally ahead of Ideogram, Flux, and Leonardo AI on image quality with its more pronounced depth of field and bokeh effect, strong detail and texture quality. Prompt adherence was similar across the four above mentioned with each capturing the scene well except that there are no artists sketching. The Designer images is not photorealistic, while Firefly’s hazy and very indistinct human figures also counted it out.
Fantasy Landscape
Winner: Midjourney and Ideogram. Midjourney’s image is the highest quality, with strong depth of field, and the best lighting effects and most detailed textures. I’m a big fan of the magical, almost cinematic atmosphere. Ideogram’s image also had a very high quality and edged out Midjourney’s slightly in relation to prompt adherence with all elements accurately depicted – the fairy houses, glowing windows, and warm sunset glow are present, and fairies are well-represented, as are the delicate bridges and ladders. The images generated by Leonardo AI and Flux were also very good but had slightly lower prompt adherence, for instance, there were no fairies in the former, and no delicate bridges or ladders for the latter. The images generated by Designer and Firefly were of noticeably lower quality and prompt adherence.
Japanese Garden Pencil Drawing
Winners: Midjourney and Ideogram. The Midjourney image came out ahead in terms of quality, with a very high level of detail, beautiful colours and contrasts, intricate textures, engaging composition. While the pencil drawing effect is visible, it leans toward a more stylised, digital interpretation of “coloured pencil drawing”. Ideogram’s output stood out in relation to prompt adherence, as the scene feels natural and peaceful, with a strong emphasis on light and shadow to enhance the serene atmosphere, while the intricate pencil line work and overall texture highly mimics the feel of a coloured pencil drawing.
Futuristic Cityscape Vector Art
Winner: Midjourney. From a quality perspective, the Midjourney image comes in first followed closely by the Firefly image. Both images are visually striking, with strong lines and vibrant colours. The former however, appears to have crisper edges and a better execution of the gradient sky. Both Midjourney and Firefly images have a good adherence to the prompt although the stylised characters are not present in the latter.
Conclusion
When it comes to quality of output (which includes the finish of the image and prompt adherence), Midjourney was the clear winner for 3 of 5 image prompts (and a joint winner for the rest). Its key strengths are its ability to create visually striking images across a variety of artistic styles, with very good prompt adherence. For those looking for the best quality images in a single package, Midjourney is tough to beat.
That being said, Leonardo AI, the newly emergent Flux, and the recently released Ideogram 2.0 are no slouches, and are surely hot on the heels of Midjourney. As compared to Leonardo AI, Flux and Ideogram appears to be able to work fluently across a broader range of styles, and could become serious contenders to Midjourney in the coming months. For now, Firefly and Designer appear to be the laggards in terms of output quality.
Whether you’re curious about AI-powered image generation or see a practical use for such tools in your day-to-day work, I would encourage you to try at least one or two of these tools to understand what they are capable of.
Finally, this article focused on the visual outputs of these tools, it's clear that choosing the right platform involves more than just aesthetics. Factors such as ease of use, cost, customisation, and even copyright considerations play a crucial role in determining which tool is the best fit for your needs. In the next article in this series, I dive deeper into these considerations, helping you to make informed decisions that align with your broader creative and business objectives.
Justin Tan is passionate about supporting organisations and teams to navigate disruptive change and towards sustainable and robust growth. He founded Evolutio Consulting in 2021 to help senior leaders to upskill and accelerate adoption of AI within their organisation through AI literacy and proficiency training, and also works with his clients to design and build bespoke AI solutions that drive growth and productivity for their businesses. If you're pondering how to harness these technologies in your business, or simply fancy a chat about the latest developments in AI, why not reach out?