AI Image Generators Face Off: Flux vs. Stable Diffusion vs. Midjourney

I play out Flux vs. Stable Diffusion in a heated five-round test to determine the best AI image generator

Stable Diffusion is a big name in the AI image generation game, so if you're not looking for popular AI chatbots like ChatGPT and Gemini to create your content for you, you might be looking at this as an alternative solution. However, there's a growing number of startups and alternatives looking to have a slice of the pie, and Flux is well worth considering for its speed, efficiency and overall effectiveness. We put together a series of five prompts designed to showcase what Flux and Stable Diffusion can do, along with a few hurdles to trip them up and uncover their weaknesses. The prompts include a familiar fictional character doing an expected action; another expected action this time being performed by real humans; a slightly more obscure prompt with a standout adjective to steer the AI, an animal doing an unlikely action; and a broader prompt that gives the AI more freedom. Let's address the elephant in the room, because neither Flux nor Stable Diffusion was able to generate an image of the tooth fairy taking a tooth from under a pillow - exactly what we'd asked for. The two artificial intelligence systems attempted to illustrate the tooth fairy and teeth in a variety of form factors, but the specific prompt had not been handled. That being said, Stable Diffusion's second image was the best of all the outputs, so that's where our points go, even though it took the best part of 38 seconds to handle our request. It doesn't get much simpler than this. It's the sort of image we see all across the internet, from studies and reports to royalty-free image sites. Both captured the feeling of an office environment well, making it look busy and noisy, exactly how you'd expect an office to be. After a poor start, Flux and Stable Diffusion both deserve recognition for their outputs, but Flux takes the lead for its quick production - Stable Diffusion took three times as long. We were just as pleased with the results from this test, with both systems creating images based on our prompt with ease. Stable Diffusion's pair of responses didn't offer a huge variation - two totally different images or one image in half the time would've been preferred. The quicker processing time and the sharper detail of Flux's image give it first prize again. We may as well copy and paste our summary of the third test, because the fourth test rendered similar results. Flux's image combines a cartoon-like armadillo with what seems to be an image of a riverbed. We prefer the results from Stable Diffusion, but it's a shame that it took an eyewatering 40 seconds. Waiting for Stable Diffusion to generate two images was painfully long, but it ultimately yielded better results, showing humans walking among dinosaurs - exactly what we asked for. Flux generated a questionable image, with a miniature headless dinosaur included, but it didn't include any humans to illustrate how it would be to walk among the ancient giants. It was much quicker, but ultimately, failed the task. Although Stable Diffusion proved to be significantly slower than Flux, it generated more accurate images more of the time, leading to better results overall. That's why, with a score of 3:2, it's the winner of our showdown between these two AI image generators. Flux is a much more efficient solution in terms of time usage, but don't think cutting corners is worth a relatively small time saving, int eh grand scheme of how much time you're likely to be spending on your entire project. Do bear in mind that Flux is a more complete solution that gives you access to video generation tools, too.

I find out which is best in the ultimate AI image generation head-to-head - Flux vs. Midjourney

In the AI era, it seems that all we ever hear about is ChatGPT, and occasionally Gemini or Apple Intelligence will steal the limelight. Beneath the headlines are hundreds of other, lesser-known AI tools that can help with all manner of tasks. Flux and Midjourney are two strong choices in the world of AI image generation, but which is best? We curated a five-round test to compare the image-generation capabilities of these systems, showcasing what they can do well and what they're not so suited to. The prompts include a familiar fictional character doing an expected action; another expected action this time being performed by real humans; a slightly more obscure prompt with a standout adjective to steer the AI, an animal doing an unlikely action; and a broader prompt that gives the AI more freedom. Nobody knows what Jack Frost looks like, yet he's spoken about every year as the nights get shorter. We asked AI to produce an image of this fictional character befitting of the winter, and the results impress. Although all five of the images produced fit the bill, we prefer Midjourney's results because they include snow and wind, adding depth to an otherwise serene image. Flux generated the most familiar image - a team of football players huddled on the pitch with an audience of spectators watching on. There's something not quite right about this image though - wrongly positioned feet and the proximity of the players detracts from the image. Midjourney created four slightly better images, but they're all in a similar painted style, lacking any clear definition. Flux's content seems fine on the surface, but look a little closer and you'll spot the headless dog in the puddle or the three-legged, cross-boddied lion-like dog to its right. Other anomalies make the output disappointing, but the apparent cat to the left of the puddle would certainly add to the chaotic feeling. Maybe Midjourney's illustration of chaos is better - the four images show dogs everywhere, and are much closer to what we'd imagined when setting out the prompt. Flux's attempt at a cheetah playing hockey, albeit ice hockey, is admirable, and the fact that it's standing on its hind legs like a human adds to the fun. However, the cheetahs produced by Midjourney are much more accurate, leading to better overall results. The fourth cheetah doesn't seem to be holding the hockey stick in quite the right place, but we can overlook that for the quality of the first three outputs. We already tried this test in another comparison between two AI image generators, but we weren't too impressed with the results so hoped for a better show this time around. The detail in Flux's images is exceptional, however it missed the mark when showing us what it would be like for a human to live with the dinosaur in question. Midjourney nailed the task, showing humans and dinosaurs interacting in a variety of settings. It didn't take too much longer to generate the output, either. We think 20 seconds is a reasonable time to wait for such a solid effort. Unfortunately, Flux ended the head-to-head with zero points in the bag, but that doesn't instantly rule it out as a worthy AI image generator. The truth is that Midjourney just outperformed Flux across all five rounds, but Flux was still able to product good images. Some just missed the mark and didn't quite respond to our prompts. On the flip side, Midjourney was consistently impressive and offered a good range of images to pick from. Generating four completely separate images adds to its appeal, but it's strictly an image generator, whereas Flux does offer some other artificial intelligence tools too, like a video generator and more, making it a good pick for social media marketers.