3 top image generation tools

And the one to watch!

Hi!

As with most areas of AI, text-to-image technology has seen rapid development, with new tools being launched and existing companies competing to innovate and outperform each other.

This week brings a few announcements that show the progress made and make this technology more accessible to a wider audience. 

Ideogram.ai only came on the scene a year ago but their launch of Ideogram 2.0 this week has shown that they’re already a contender for the top spot with some suggesting it rivals Midjourney, one of the leading AI image generation tools.

Midjourney has news of its own with the announcement of a limited free trial giving everyone access to their powerful tool. 

So let's take a look at some of the recent developments and examine how they're shaping the landscape of AI image generation.

Past challenges

There are some key areas where AI-generated images have fallen short in the past:

  • Realism 
    Early AI-generated images of humans were clearly fake - hands with too many fingers, hair that looked like cotton candy or elements that almost melted into each other

  • Text 
    Many AI image generators have struggled to incorporate readable text within an image, instead using approximations of real words or in some cases, plain gibberish

  • Prompt adherence 
    Following a prompt fully is another area where some tools have fared better than others. Where detailed requirements are set out in a prompt, many tools will fulfill part of the brief but not adhere fully

But the teams behind these tools have worked hard to address the tell-tale signs and challenges and the resulting images have become increasingly convincing and versatile.

Editing options have also been developed providing more customization options and giving users the ability to fine tune images to their exact needs/requirements. 

Let’s take a look at 3 key players.

Ideogram

The latest release from Ideogram is a big improvement on the first version and they seem to have nailed the incorporation of text within an image. This is particularly useful for anyone needing to create logos, lettering on signs or any other text-based elements within their AI-generated images.

Beyond the impressive performance of the tool itself, the accessibility for users is also a big advantage. The free plan has a generous offering of 10 slow credits per day which translates to a potential 40 images.

Any images you create on the free plan are public by default and you’d need to switch to a paid subscription for priority credits and premium features but for most hobby and non-professional users, the free plan should be more than enough to meet their needs. 


Midjourney

Initially only available through Discord, Midjourney created a website last year to improve access and the free trial announced this week means everyone gets a chance to test out this much-praised tool’s capabilities.

When it comes to realism, Midjourney is seen as a superior option to many other tools and the output is consistently high quality. An image created using Midjourney of the pope wearing a white puffer jacket went viral last year with many people failing to spot it as an AI fake due to its realistic rendering.

Midjourney isn’t quite as good as Ideogram for incorporating text in an image and can deviate from the prompt at times. However, the image editor allows you to refine your chosen image and change specific areas so you can make precise adjustments to get just what you need.

The limited trial allows users to create up to 25 images, after which you’d need to opt for a paid plan.

Sponsored by Penfriend.ai

Penfriend generates high quality SEO articles, without prompting.

Get in-depth, 3,000+ word articles designed to rank in Google:

  • in 3 clicks

  • in 15 minutes

  • and for as low as $18

DALL·E 

Created by OpenAI, DALL·E was probably the first mainstream image generator and it continues to be widely used and respected. 

It might not be quite as advanced as some of the other image generation tools when it comes to realism but it excels in prompt adherence. While other tools seem to pick and choose which elements from the prompt to keep, DALL·E 3 does a great job of following prompts to the letter and incorporating all instructions and elements.  

Access to DALL·E 3 was previously limited to paid ChatGPT Plus users. However, now free ChatGPT users can create 2 images per day. For those needing more images, Microsoft Designer's Image Creator, which is powered by DALL·E 3, offers a free-to-use alternative with no daily limit.

The one to watch - FLUX 

FLUX.1 from Black Forest Labs should definitely be on your radar. It’s had a good reception so far and performs well creating realistic images of humans.

As it’s open source, development could be rapid and with the team working on a text-to-video model it will be interesting to see how this technology develops. Elon Musk is already a fan, choosing FLUX as the foundation model for his X AI tool, Grok.

A squirrel wearing a monocle and checked flat cap, sitting on a log and reading a newspaper with a cup of tea balancing on a table made out of an acorn nearby

Prompting tips

As with content generation tools, perfecting your prompts is key to getting good results. If you’re struggling with your prompts, consider the following tips:

  • Detail
    The more information you give in your prompt, the better the chance of the outcome matching your expectations.

    Think about style, colors, setting and composition as well as subject matter.

  • Be specific
    Using strong visual keywords should help to create the desired outcome so choose vivid, descriptive language.

    For example, ‘a towering oak with gnarled branches’ rather than simply ‘a tree’.

  • Negative prompts
    Some tools allow you to include negative prompts to cover anything you don’t want in the image.

    For example, you might want to specify ‘no people’ for a city street scene at night.

  • Experiment
    Not all tools allow you to edit or refine results so you may need to have multiple attempts to get the desired outcome.

    With each iteration you can see what works and what doesn’t until you’ve perfected your prompts.

  • Ask for help - Ask Claude, ChatGPT or Gemini for help improving your prompts or use a tool like Glif which can incorporate prompt optimization into a workflow. 


With these tools at your disposal you can let your creativity run wild. Experiment with each one and with a bit of trial and error, you might just find the perfect fit for your creative vision.

OK, speak soon!

Inge
Top Apps Founder

P.S. Check out the Top Apps AI courses here.