Nano Banana 2.1 Comes Close to ChatGPT in Image Tests

TechRadar’s Eric Hal Schwartz compared Google’s Nano Banana 2.1 with ChatGPT’s Images 2.5 using identical prompts. The tests covered reflections and lighting, poster text, character consistency and following detailed instructions. The results were close, though the generators often interpreted the same prompt differently.

For a glass snow globe containing a Victorian greenhouse during a thunderstorm, both tools made detailed images. Schwartz found Nano Banana 2.1’s globe reflections convincing and its lighting somewhat more natural, while noting that its wet tabletop seemed odd for an indoor scene. ChatGPT’s version featured a golden-lit greenhouse inside a large glass sphere.

A second prompt asked for a retro-futuristic travel poster for Saturn’s moon Titan with the exact text “TITAN IS CALLING,” “Seven nights beneath the rings,” and “Starting at $2,499.” The source excerpt does not report the results of that test.

In a four-panel comic prompt, Nano Banana 2.1 kept an orange fox and its green scarf consistent across the story. Google’s panels read more like one connected story, while ChatGPT’s images were more elaborate but varied somewhat in the fox’s proportions and presentation.

For the final prompt, both generators included a dragon notebook, a north-pointing compass, a pocket watch showing 3:15, exactly three red wax seals and a blue gemstone in an overhead desk image. The source excerpt ends before describing the full comparison. Overall, Schwartz found the tests surprisingly close, with different strengths across the examples.