A year ago, when you wanted an image for a blog post, you'd spend hours searching stock photo sites. Now you can write a few sentences and generate the image you want in seconds. AI image generation technology has advanced at an incredible pace over the last two years, and it's now used even in professional projects.
Over the last six months, we've used three different AI image generation tools on dozens of projects. We've produced visuals for blog posts, social media content, presentation slides and even client presentations. We've discovered each tool's characteristic style and where it does best.
Midjourney: Artistic Quality
Midjourney is the undisputed quality leader in the world of AI image generation. The images it produces are the most aesthetically impressive. It's unrivaled especially for artistic, atmospheric and concept images.
We first used Midjourney in a client's brand identity work. We needed to create different visual concepts for the brand, and with the traditional method we'd have spent hours on each concept. With Midjourney, we created 20 different concepts within two hours. When the client saw these images in the meeting, they were very impressed and even wanted to use some of them directly as social media content.
Midjourney's strengths:
- Photorealistic and artistic quality
- Consistent aesthetics and a stylized look
- The community is very valuable as a source of inspiration
- Keeps improving with version updates
However, Midjourney runs through Discord, which may feel odd to some users. It also still struggles with images that contain text, and it doesn't fully understand Turkish prompts. You need to write in English.
A designer friend of mine canceled their stock photo subscription after discovering Midjourney. Now they produce original images that fit their clients' needs exactly. Finding a suitable image for a blog post used to take them half an hour; now they get it done in five minutes.
DALL-E: OpenAI's Image Master
DALL-E is the most accessible AI image tool, integrated right into ChatGPT. If you have a ChatGPT Plus subscription, you can use it at no extra cost.
DALL-E's biggest advantage is its natural language understanding. You can give complex, detailed descriptions and generally get results close to what you wanted. The ability to select and change a specific part of an image is also very handy. For example, in an image you've created, you can change the background color or the position of an object.
We use DALL-E especially for social media visuals. Being fast and practical is a big advantage when producing content frequently. We can create all the visuals for a week's social media calendar in one sitting.
On the downside, DALL-E's aesthetic quality lags behind Midjourney's. The difference is especially noticeable in human faces and detailed scenes. There's also a daily usage limit.