Google has rolled out Nano Banana 2.1, the latest version of its AI image generator, inside Google Gemini, and early side-by-side tests against OpenAI's ChatGPT image generator, Images 2.5, show the two American tech giants are now nearly level in quality. For Zimbabwe, the message is clear: the battle over who controls the world's digital imagery is being fought entirely on foreign soil, and our nation must build its own capacity rather than remain a consumer of other people's innovations.
The test, published by TechRadar on 13 October 2026, used identical prompts on both platforms and found the results surprisingly close, with each generator showing different artistic instincts but neither committing the glaring visual errors that once made AI images instantly recognizable.
What is Nano Banana 2.1?
Nano Banana 2.1 is Google's upgraded image generation model, now the default inside Gemini for subscribers of Google AI Plus, Pro or Ultra. Free users remain on the older Nano Banana 2 model. Google promises sharper details, better instruction-following, more accurate text and greater consistency across images.
Its main rival is Images 2.5, the image generator built into ChatGPT by OpenAI, which has made considerable progress with realistic lighting, complicated compositions and text rendering.
How did the two AI generators compare?
The tester ran four identical prompts through both systems, placing ChatGPT's results on the left and Gemini's on the right.
Light, glass and reflections
A prompt for a photorealistic Victorian greenhouse inside a glass snow globe during a thunderstorm produced beautiful images from both. ChatGPT delivered an elaborate, golden-lit greenhouse on a polished wooden base, while Nano Banana 2.1 produced a more restrained miniature with dark metal framing and particularly convincing reflections on the globe.
Text rendering
Text has historically been an embarrassing weakness of AI image generators. Asked to design a retro-futuristic travel poster for Saturn's moon Titan with the exact text