I tried GPT-4o and it's a gamechanger for free ChatGPT users

We’ve known that OpenAI is actively developing its next-gen GPT-5 language model for some time now. But earlier this week, the company surprised us with an incremental update to GPT-4 Turbo. Dubbed GPT-4o, the new model is a massive update that brings real-time voice interaction, video capabilities, and faster and more accurate responses. But flashy demos aside, GPT-4o also marks the first biggest update for free ChatGPT users in well over a year. Here’s why it’s such a big deal.

What is ChatGPT-4o?

ChatGPT-4o or GPT-4o is OpenAI’s latest and most advanced language model that can reason across audio, video, and text in real-time. The model can respond to audio inputs within an average of just 320 milliseconds, about as fast as human response time, making it capable of natural conversation.

Last year, I wrote about ChatGPT’s voice conversation feature and how it already mimicked a real-life conversation almost perfectly. GPT-4o elevates that experience even further. While the previous feature used different models for speech recognition and response generation, GPT-4o has been trained to handle the whole process from end-to-end. Armed with the capabilities of a large language model like GPT-4, OpenAI says that the AI can even detect emotion in voices.

GPT-4o’s improved voice capability would be a big deal on its own, but that’s not the only new feature OpenAI showed off. The new language model also offers significantly faster responses, thanks to a new tokenizer. In non-English languages, and especially ones that don’t use the Latin alphabet, OpenAI says the new model will feel twice as fast as the previous GPT-4 Turbo model.

The bombshell announcement of the event was that GPT-4o is coming to all ChatGPT users. That’s right — you can access the latest and most capable version of the chatbot for free, no ChatGPT Plus subscription necessary. This marks the first-ever accuracy and speed update for free users since the chatbot’s release all the way back in late 2022.

The good news for free users doesn’t end there. Since GPT-4o is an extension of OpenAI’s GPT-4 language model, it’s also capable of searching the internet and fact-checking its responses. Until now, free users were stuck with the GPT-3.5 model that was significantly less intelligent and had a knowledge cut-off of January 2022.

GPT-4o in action: How to use it

If you haven’t logged into ChatGPT in a while, now’s the time to change that. OpenAI says that it’s opening up access to GPT-4o to all users in waves, effective immediately. I’ve been using Microsoft Copilot for about a year as it offers free access to GPT-4 and, consequently, rarely ever used ChatGPT in that time. Still, I was greeted with access to GPT-4o in a “limited capacity” as soon as I logged in.

It’s worth noting that GPT-4o’s voice conversations feature isn’t available yet. OpenAI CEO Sam Altman said that it will begin rolling out in the coming weeks. Voice conversations will also be exclusively available to ChatGPT Plus users, at least for the foreseeable future.

With that disclaimer out of the way, GPT-4o is still an impressive update in its own right — especially for non-paying users. Type in a prompt and you’ll notice that the model responds impressively quickly in comparison to GPT-4. However, it may not necessarily be as fast as the legacy GPT-3.5 model which has been relentlessly fine-tuned and optimized since its initial release four years ago.

Testing GPT-4o: Is it really more accurate?

Speed aside, GPT-4o’s big highlight is that it delivers significantly more accurate responses. And unlike GPT-3.5, it doesn’t have a hard knowledge cutoff date either. This means you can finally ask the chatbot factual questions without having to worry about it hallucinating or spitting out false responses. Just how much of an upgrade is GPT-4o? Quite a big one in my testing. Here’s a side-by-side comparison where I ask both GPT-3.5 and GPT-4o a simple factual question:

As you can see from this relatively simple example, both language models deliver the correct response. However, GPT-4o was significantly more confident in its response and provided a detailed answer. The older GPT-3.5 model (which was the only model available to free ChatGPT users until now) responded from memory instead, which explains why it asked us to verify the information with an official source.

Interestingly, I could not switch from GPT-4o to GPT-3.5 for this particular question as the option was grayed out with the message “This model doesn’t support using tools”. So if you ask ChatGPT a factual question, it will now automatically browse the internet and use the latest model. Here’s another example where I asked ChatGPT which chip the Galaxy S24 Ultra contains, with the condition that I’ll be buying it in the UK.

Thanks to GPT-4o’s ability to search the internet, it correctly said that the S24 Ultra comes with the Snapdragon 8 Gen 3 worldwide. It even clarified that the regular and Plus S24 variants will come with the Exynos 2400 in Europe. I’ve asked other chatbots the same question in the past only to get a confident-sounding but inaccurate response. Just look at the response from GPT-3.5 above.

While we’re still a few weeks away from experiencing the full potential of GPT-4o, namely its voice capabilities, the baseline experience is still impressive in its own right. Better yet, it’s universally accessible to everyone for free. So if you’ve been using Microsoft Copilot or Google Gemini instead of ChatGPT, now might be the time to switch back.