By skyforbes First, I tested it for coding. In the past, my favorite tool was the Codex CLI because it was reliable, unlike Claude, which often made mistakes. I took my application—which uses an HTML and JavaScript frontend with a Python backend—and simply told the AI to improve the GUI. It genuinely improved the user interface, and everything worked perfectly on the first run with no errors. I used the Gemini CLI for this, and I believe it is a huge step forward for programmers.
Next, I tried creative writing. I took the sci-fi story *Ask a Foolish Question* by Robert Sheckley and asked Gemini to write a sequel. The result was amazing; it handles creative writing very well.
Finally, I tested its image capabilities. I had a couple of images of handwritten text from over 200 years ago that even I couldn't read completely. I uploaded them and asked the model to transcribe the text, and it provided a complete and accurate result.
I am more than impressed with this model and will definitely continue playing with it.
Thank you guys for this great product.