AI image generators are all the rage today, but which ones do people favour and why? It’s often, if not usually, to do with the sort of result you want to see, because they all have their own characteristics and idiosyncrasies, which are mostly due to the what they’ve been trained on. Here’s a round-up of the latest runners and riders in the race for AI image supremacy.
DALL-E 3
CNET’s 2024 Editor’s Choice award winner. Versatile, customisable and unique in its field for its conversational flow. Good at receiving follow-up requests to requests and able to handle complex queries. Available as part of a $20/month ChatGPT subscription.
Pro’s – understands long, complex queries, generates engaging dynamic images, conversational style allows for easy modifications
Con’s – photorealistic images can look fake, slow to generate images
Adobe Firefly
Built directly into Adobe Creative Cloud, so great for pro’s with an account who want to experiment. Refined, packed with options and trained on Adobe’s own image catalogue, which includes high-quality licensed and public domain content.
Pro’s – excellent artistic styles, good fine-tuning options, competitive generation speed
Con’s – struggles with photorealistic images and elements that contact or interact, doesn’t accept complex queries
Leonardo
Great for those on a budget. Fast, and the comprehensive free plan provides ample generation credits. Post-generation editing tools sit behind a paywall, but there is a free prompt improvement tool and other customisation tools.
Pro’s – clear, detailed images, helpful prompt engineering tools
Con’s – No post-generation editing with the free plan, unsatisfactory security policy
Canva
Magic Media is the minimalist AI image generation part of this user-friendly platform. It’s good for budget-conscious amateurs, makes it easy to integrate what it generates into other projects and has a robust security policy.
Pro’s – good for fun, whimsical, cartoon-y images, desktop and mobile versions are both easy to use, doesn’t share your content to train its model
Con’s – a hard limit on how many images you can generate with a free plan, difficult to generate anything beyond a square image
Stable Diffusion
Open source, great at creating detailed and colourful AI images and one of the most commonly used tools around. Lots of editing tools, if you’re a Stability AI user.
Pro’s – fast generation time, great editing tools, very creative
Con’s – complicated availability, product images are too realistic
Source: CNET