You know how they say: “It’s not a competition!” Well, don’t let them lie to you; Everything is a competition, especially when it comes to AI. There is rarely a day when I am not testing AI capabilities between various chatbots, and I am almost always surprised by the results. Some platforms are actually better than others, at least for some tasks.
This journey started with Notes on my iPhone 17 Pro Max. I typically like to record interviews on an Android smartphone like the Google Pixel 10 Pro Fold, where the fantastic Recorder app expertly captures each utterance and, in transcription, does a deft job of separating and labeling each speaker.
Gemini 3 Pro puts on the gloves
Over the past few months, I’ve been impressed with the Google Gemini’s capabilities, especially the latest 3 Pro models, and how it seems to handle almost any immediate request with aplomb.
Now that I had the idea, I had to figure out how to get Gemini to listen to the recording. Playing the audio through my iPhone’s speakers and asking Gemini to listen was out of the question because I was concerned about how well, say, my desktop microphones could pick up the sound coming out of the iPhone’s speakers. Also, I was in the office and didn’t want people to hear the private conversation (until I posted a story).
First, I discovered that you could download the audio file from Notes. In playback, below the three dots, there is a Share button that allows me to airdrop the audio file to my 14-inch MacBook Pro. It comes as an MPEG-4 (M4A) file.
Back on Gemini 3 Pro, I selected the “+” sign in the prompt field, chose the M4A audio file, and added this short message: “Listen to this, transcribe it, and make sure you identify the different speakers.”
There was no back and forth. Germini 3 Pro quickly began spitting out the full transcript with the speakers identified as “Interviewer” and my subject’s name and title. It’s worth noting here that this is the only thing the Gemini 3 Pro inexplicably got completely wrong. Although my subject spelled his name at the end of the talk, Gemini chose a different one. Other than that, however, Gemini perfectly identified when it was me or the subject speaking. And the precision was really impressive.
For completeness, I asked Gemini 3 Pro to correct my subject’s ID and list me as “interviewer.” With that taken care of, I happily used the transcript to help drive my full story.
In this corner, ChatGPT
However, I was naturally curious if ChatGPT 5.1 (with a Plus account) could perform the same task.
In the ChatGPT message window, I selected the audio file and entered the exact same message. ChatGPT told me, “I can definitely transcribe audio, but I can’t access or play the .m4a file directly from the location you referenced.”
What followed was a lengthy exchange in which ChatGPT kept suggesting different ways to upload the file, including transforming it into a zip file. No matter what I did, ChatGPT showed the audio file in the prompt window, but I couldn’t listen to it.
It seems in this little competition, the Gemini 3 Pro is the winner, turning a frustrating problem into an easy victory. The less said about how useless Apple’s Notes transcription is, the better.
The best business laptops for every budget
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
