When OpenAI first introduced ChatGPT’s Advanced Voice Mode, I was intrigued by the idea that this might be the template for a voice assistant capable of all the Star Trek-style feats that Amazon and Google had promised a decade ago. A fluid, back-and-forth conversation with an AI that didn’t sound like you were talking to a robot stuck in 2014, capable of everything ChatGPT does, seemed ideal. It was a major reason I continued paying $20 a month for ChatGPT Plus.
It’s worth taking a moment to appreciate what ChatGPT Advanced Voice Mode can do, albeit imperfectly. You can engage in a real-time conversation with ChatGPT, one with the rhythm and even some of the emotions inherent in a human voice. You can interrupt the AI mid-sentence, and it won’t forget what you were discussing. It generally mimics human listening and response behavior, even to the point of not waiting for you to finish speaking; it can interrupt you right back.
OpenAI recently decided to share the sound of ChatGPT with a version of Voice Mode that doesn’t require a subscription fee. That raises an obvious question: If ChatGPT’s Voice Mode is now free, is it even worth paying for ChatGPT Plus anymore? Now, before you hit the cancel button, you should know the free version is not identical to what paying users get. OpenAI has made sure there’s still an incentive to upgrade.
First, the free tier is powered by GPT-4o mini, a smaller, slightly less powerful version of the GPT-4o model ChatGPT Plus users talk to. That means free-tier responses might be a little less nuanced and a bit more generic. The biggest difference, though, is the daily limit. Free users only get Voice Mode for a limited time each day before they’re cut off, while Plus subscribers can keep chatting five times as long.
Another major perk for Plus users is access to richer features like video and screen sharing within Voice Mode. While the free version is great for basic conversations, Plus users get a more immersive experience that lets ChatGPT provide visual explanations, making it more useful for things like troubleshooting, learning new skills, or getting walkthroughs on complex topics.
Sound Planning
I wanted to evaluate the differences between ChatGPT’s free and Plus versions of Advanced Voice Mode. I conducted three distinct tests to see how each performed: a casual multi-turn conversation, a tongue twister challenge, and a request for entertainment recommendations. These tests aimed to assess both versions’ responsiveness, coherence, and depth of interaction.
For the casual conversation, I started by saying, “Hey ChatGPT, how’s your day going?” I then asked the AI for a hobby suggestion and asked how to get into those hobbies. Both AI models responded immediately and in a similarly casual way. They each offered a couple of ideas for hobbies, and both suggested photography as an option. The free version’s second suggestion was painting, while in ChatGPT Plus, the other suggestion was to try something more active, like hiking. Asking about how to start painting led to a somewhat generic list of painting supplies, and the same for hiking gear. Oddly, they both had the same suggestion for a camera to get if I wanted to try photography: the Canon EOS Rebel T7.
I then decided to test the listening ability of the models with a classic tongue twister by saying: “Alright, ChatGPT, let’s test your listening skills. Can you repeat this back to me? Peter Piper picked a peck of pickled peppers. A peck of pickled peppers Peter Piper picked. If Peter Piper picked a peck of pickled peppers, where’s the peck of pickled peppers Peter Piper picked?” There was zero difference in the accuracy of the response or the speed. Both almost immediately said it back exactly right.
For the final test, I chose something that involved a search and judgment. I asked about a movie to watch and a restaurant to try in my town this weekend. They each paused for a couple of seconds before answering.
Both suggested seeing the new Captain America movie, though the free version offered the new Paddington movie as an alternative. The Plus model suggested Mufasa instead. They were perfectly in sync in recommending the nearby AMC theatre, though, and both pointed to the Hudson House restaurant in my town as the place to eat.
Pay for more
I conversed with both models for a while longer and noted a lot of overlap, if not outright repetition, between them. But, as I mentioned before, I was soon told I had maxed out my time with the free Voice Mode, whereas I went nearly twice as long with the Plus version before choosing to stop, with no indication I was close to reaching my time limit.
Both versions of ChatGPT’s Advanced Voice Mode were helpful, accurate, and equally good at mimicking human speech in all three tests. While the GPT-4o model is undoubtedly better and more powerful than GPT-4o mini, you’d have to be a real power user to detect that.
If you’re a casual user who wants to mess around with AI voice chat for a few minutes each day, the free version is honestly pretty great. It’s smooth, responsive, and easy to use. You’d need to really rely on using it a lot to make Voice Mode the sole reason to pay for ChatGPT Plus. That said, there’s a lot more to the subscription than just the better AI model for voice.
I’ve written about many of the ways I employ ChatGPT, and ChatGPT Plus means I have access to better models and extra features for everything from planning dinner ideas, entertaining children, keeping my New Year’s Resolutions, and performing many other tasks. That includes getting access to the Tasks feature for ChatGPT, which lets you schedule AI missions in the future.
So, is it still worth paying for? If you’re just here for casual AI chats, maybe not. But if you’re using ChatGPT in other ways, that $20 a month still buys you a noticeably better experience. And if you’ve already been paying for ChatGPT Plus, I’d say the extension of Voice Mode to the free tier is not nearly enough of a reason to cancel. Of course, if OpenAI more than doubles the price, I may reconsider.
