Google has started testing Gemini's next big upgrade for working with images

Google has already given us lots of great tools for working with imagery, and a couple months back we spotted evidence of development on another big one: Letting you mark up images you share with the Gemini app. Then a few weeks ago a leak revealed that Google was prepping the same sort of markup trick for Gemini on the web. And now, both appear to actively be in testing.

After getting a tip from user Diego on Telegram, we fired up Gemini to check, and can confirm we’re also seeing this markup ability present in both the Gemini app, as well as Gemini on desktop in Chrome. Upon first attaching an image in Gemini on Android, the app will let you know about the new ability:

We spotted that same editing-based intro note in our last teardown, but rest assured, that’s not the only thing this tool is good for, and just like we initially expected, it also works for directing Gemini’s attention to certain areas of an image.

As you can see with both image editing as well as analysis, Gemini is easily able to understand the part of the input picture we’re concerned with when we enter our prompt.

The editing is probably the more successful demo here, but even as Gemini’s analysis on the right missed the mark (HOW are you going to confuse Mantzoukas and Scheer?), it clearly understood we were asking about the left-most of the three pictured individuals.

Right now we still can’t say how widespread this testing is, or if Google may actually be in the process of just full-on rolling it out. Hopefully we’ll have a better sense of availability after the weekend, but go ahead and check your devices now to see if you can start taking advantage of these tools.