AI makes pictures, but can it give The Rock bangs? Five bots take our image challenge.

Review by Geoffrey A. Fowler

and

Kevin Schaul

December 16, 2025 at 5:00 a.m. ESTToday at 5:00 a.m. EST

TapClick to start Go to Next Page

6 min

Lots of artificial intelligence tools can make images. But which is the best?

We challenged five leading AI image generators to edit photos, convey emotion and make art. Tap through our tests and see how they ranked with our panel of expert judges.

About this story

We conducted the tests in November 2025 using Adobe Firefly Image 5 (beta), ByteDance Seedream Image 4.0, Google Gemini 3 Nano Banana Pro, Meta AI and OpenAI’s ChatGPT 5.

We ran the same prompt through each AI tool, collecting the first image each bot created. Judges scored each image on a 10-point scale without knowing which AI tool had created it.

The original image of Dwayne Johnson was created by David Shinbone. The original image of Kristen Stewart and Robert Pattinson was submitted by accidentalpaparazzi. Both images were licensed under Creative Commons 2.0 and modified using AI with the permission granted by the licensing agreement.

Photo of Carson by Ava Carson, photo of Dreszer by Dahlia Dreszer and photo of Naik by Angi Lewis.

Geoffrey A. FowlerGeoffrey A. Fowler is The Washington Post’s technology columnist based in San Francisco. He joined The Post in 2017 after 16 years with the Wall Street Journal. He won the 2020 Gerald Loeb Award for commentary.@geoffreyfowler

Kevin SchaulKevin Schaul covers artificial intelligence for The Washington Post. He uses data, visuals and code to examine how AI actually works and its real-world implications.@kevinschaul

More For You

Pentagon plan calls for major power shifts within U.S. military

December 16, 2025 at 9:08 a.m.

She was MAGA’s favorite conspiracist. Then she dug into Charlie Kirk’s killing.

December 16, 2025 at 9:12 p.m.

What a scientist who studies protein and healthy aging eats in a day

December 16, 2025 at 7:00 p.m.