gleitfrosch/Getty
Before ChatGPT, there was DALL-E — the first iteration of OpenAI’s now-famous text-to-image AI model. It was released in January 2021, giving the world an early taste of the future of AI-generated imagery that was soon to come.
Today, there’s a long roster of image-generating AI tools to choose from. Companies like OpenAI, Google, Midjourney, Stability AI, Ideogram, and Black Forest Labs all offer their own models, each with its own particular strengths and weaknesses.
(Disclosure: Ziff Davis, ZDNET’s parent company, filed an April 2025 lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
Also: The best AI image generators of 2025: Gemini, ChatGPT, Midjourney, and more
Now, a new website allows you to keep track of — and vote on — the leaders of this constantly evolving (and increasingly crowded) category of AI models.
How it works
Launched in beta last month, LMArena.ai ranks AI models “based on their ability to generate images based on text descriptions,” according to its homepage. The website was originally founded as a research initiative through UC Berkeley with the name Chatbot Arena, and has since become a full-fledged platform, with financial support from UC Berkeley, a16z, Sequoia Capital, and elsewhere.
Also: I test AI tools for a living. Here are 3 image generators I actually use and how
Every user’s vote impacts models’ rankings on the scoreboard through the Elo system, which was originally devised to rank players in chess and other competitive games. Model names remain anonymous to reduce bias; only after a vote has been submitted is the name of the model revealed to the voter.
At the time of this writing, a little under 281,000 votes have been submitted on the performance of nine text-to-image models. OpenAI’s GPT Image 1 currently holds the top spot, followed by Google’s Imagen and Luma’s Photon.
Screenshot by Sabrina Ortiz/ZDNET
Beneath the scoreboard, the website depicts a series of charts that provide greater detail into how each model’s ranking is determined.
Screenshot by Sabrina Ortiz/ZDNET
Separating the wheat from the chaff
The launch of ChatGPT in November 2022 sparked a flurry of similar generative AI tools as companies vied for dominance in this new product category, which some forecast could be the most lucrative since the smartphone or even the internet.
Also: ChatGPT’s new image generator shattered my expectations – and now it’s free to try
As a result, users are faced with a massive number of choices. Every generative AI model has unique capabilities, pricing structure, and data privacy policies, making it difficult to know which ones to turn to. Websites like LMArena have stepped in to provide some clarity.
AI developers have a lot to gain from these websites, too, as they provide platforms for sharing valuable feedback and are based on extensive testing. This leads to a steady stream of use that, in turn, helps the models themselves improve. ZDNET went hands-on with dozens of AI image generators and identified the top 10, scoring them on different features and attributes such as cost, guardrails, speed, quality, and more.
Get the morning’s top stories in your inbox each day with our Tech Today newsletter.
