We Tried Out GPT-4o, And It

If you’re curious how much faster GPT-4o is, the difference between it and GPT-4 is staggering.

Key Takeaways

GPT-4o is a game-changer with its incredible speed and improved language support.
The new model is significantly faster and cheaper for developers to implement.
Users are encouraged to give GPT-4o a try for a more efficient generative AI experience.

Remember when comparing Google’s Gemini to ChatGPT just how much faster Gemini was? OpenAI clearly didn’t like that, because GPT-4o is so much faster than GPT-4 ever was that it makes GPT-4 look archaic in comparison. We’ve been testing out GPT-4o for a little while since it launched, and the difference is staggering. As you can see even in the above video, asking GPT-4o to generate a CSV took less than a minute, whereas GPT-4 took nearly as long to just generate the cities being used in the example.

GPT-4o just launched earlier today, and with it came a number of improvements. It’s so fast that you can converse with it out loud, and it’s significantly cheaper for developers to use in their apps, too. It’s overall just a better model than GPT-4, and I’m so excited to play with it more over time.

OpenAI GPT-4o improvements, showing 2x faster, 50% cheaper, and 5x higher rate limits

What is GPT-4o? Everything you need to know about the new OpenAI model that everyone can use for free

GPT-4o has just been announced, and it’s a faster, better successor to GPT-4 Turbo that everyone can use. Here’s everything you need to know!

GPT-4o is the big GPT-4 improvement we’ve wanted for a long time

It’s just that much faster

GPT-4 has been more than good enough for a lot of people for a long time now, but the biggest downside of it was how slow it was in comparison to other LLMs. Google’s Gemini absolutely smoked ChatGPT in output speed, and the same with Claude 3 from Anthropic. GPT-4 still had some other advantages that kept people coming to it, but the speed, quality of answers, and supported languages are what really held it back.

In the above video, GPT-4o generated its 488-word answer in less than 12 seconds. A similar response would require nearly a minute of generation under GPT-4 at times, and that’s also with the slight improvements that GPT-4 Turbo (and thus, GPT-4o) brings as well. It’s shockingly fast, and alone is enough to question whether ChatGPT is the king of LLMs yet again.

If GPT-4’s biggest downside to you was its generation speed, rest assured that that problem is gone entirely now. It’s so much faster than it once was, and it’s hard to understate just how much of a difference it makes. I’m pretty impatient at the best of times, and sitting there waiting for GPT-4 to generate responses was a killer whenever I used it. This makes it so much more usable for plenty of different use cases, and I highly recommend trying it out. Even as a free experience, it will be one of the best LLMs around.

ChatGPT vs Microsoft Copilot vs Google Gemini: What are the differences?

If you’ve been trying to figure out which generative AI tool is better, you’ve come to the right place

Other improvements GPT-4o brings

GPT-4o builds on GPT-4 Turbo

Source: OpenAI

GPT-4o is significantly faster than regular GPT-4, but even twice as fast as GPT-4 Turbo. Plus, it’s 50% cheaper for developers to implement, and has much higher rate limits to use it, too. It benefits from GPT-4 Turbo’s improved reasoning capabilities, and GPT-4o even has improved language support in languages such as Korean, Russian, Arabic, and more.

If you want to give it a try and see for yourself, OpenAI is rolling out GPT-4o support to people’s accounts starting today. They said that it will be rolled out over the next few weeks, but be sure to check your account to see if you have it. I received it quite early, but I’m currently in the United States. Other regions may need to wait longer, and you may also increase your chances of getting it sooner with a ChatGPT Plus subscription.

All of the responses GPT-4o has been giving me have been this fast. It’s not just that these prompts have been quick, it’s literally any prompt. OpenAI has knocked it out of the park, and given how fast it is, it’s no wonder that it works for audio generation, too. We’re excited to see where the future of GPT-4o takes us, and I’m even more excited about trying out audio prompts with GPT-4o in the future.