Google's New AI Image Maker Goes Head-to-Head with ChatGPT

Google has quietly upgraded its AI image generator, Nano Banana, to version 2.1. It's now built into Gemini, Google's answer to ChatGPT. The new version promises sharper details, better instruction-following, and more accurate text in images. But the real question is: does it beat ChatGPT's image tool, Images 2.5? A recent side-by-side test gave identical prompts to both. The results were very close, with each tool showing different strengths.

In one test, both created a photorealistic snow globe with a tiny greenhouse inside. ChatGPT's version was more elaborate and golden-lit, while Google's looked more natural and had convincing reflections. In another test, both designed a retro travel poster for Saturn's moon Titan. ChatGPT's poster looked like a luxury ad, but it had a small spacing error in the price. Google's poster was more cartoonish but got all the text right, including the price. Text has always been a weak spot for AI images, so Google's accuracy here is a big win.

Character consistency—keeping the same character looking the same across multiple images—is another challenge. The test didn't finish that part, but it's a common problem. Overall, both tools are impressive and fun to use. If you have a Gemini account, you can try Nano Banana 2.1 right now. ChatGPT's tool is also available to Plus subscribers. The competition is good news for users: it means both companies will keep improving.

So what does this mean for you? If you need AI-generated images for work or fun, you now have two strong options. Google's tool is better at text, while ChatGPT's might be more artistic. Neither is perfect, but they're getting better fast. And since both are built into chatbots you might already use, you can test them yourself without extra cost. Just don't expect perfection—AI still struggles with complex scenes and exact details.