ChatGPT 3.5 and 4.0 Compared: Which Version Is More Human?

Since its release, the artificial intelligence ChatGPT has been on everyone’s lips. Developer company OpenAI has now even introduced a paid subscription that is said to contain an improved model. But: What are the differences between ChatGPT 3.5 and 4.0? And: Which version is more human? A comparison.

The ChatGPT Plus paid subscription has also been available in Germany since mid-February 2023. Developer company OpenAI promises that the Plus subscription will deliver faster response times than the standard free version. In addition, ChatGPT Plus should have additional functions. Among other things, the AI ​​runs much more reliably in the paid subscription.

The differences between ChatGPT 3.5 and 4.0 in comparison

In addition, OpenAI recently released an update with GPT-4.0, which should increase the probability of providing more factual answers by up to 40 percent. From now on, the AI ​​can provide “human-level performance in various professional and academic challenges” and is fundamentally even more human.

GPT-4.0 is currently only included in the ChatGPT Plus paid subscription. The free standard version is still based on GPT 3.5. But: Is ChatGPT 4.0 really more human than its predecessor?

We tested both versions and gave them the same tasks. In comparison: a human-made text. The task: “Write me 100 words about the beginnings of the iPhone.”

ChatGPT 3.5:

“The iPhone was first introduced by Apple in 2007 and revolutionized the way we use our cell phones. Steve Jobs, the late co-founder of Apple, presented the iPhone as a combination cell phone, iPod and Internet communication device. It featured a touchscreen that replaced the traditional keyboard, along with a variety of features such as a web browser, camera, and music player. The iPhone quickly became a bestseller and led to a huge increase in demand for smartphones. The beginnings of the iPhone thus changed the mobile technology industry forever and continues to influence our daily lives to this day.”

ChatGPT 4.0:

“The iPhone, a revolution in the world of mobile phones, was first introduced in January 2007 by Apple founder Steve Jobs. The groundbreaking combination of mobile phone, music player and Internet communication device redefined the smartphone and ushered in an era of mobile technology. The first iPhone featured a capacitive multi-touch display, an accelerometer to detect device movement, and an intuitive user interface that leveraged the iOS operating system. With its innovative features and beautiful design, the iPhone fundamentally changed the mobile industry and set a new benchmark for technology and user experience. This success led to the development of a number of models and cemented Apple’s position as one of the world’s leading technology giants.”

ChatGPT 3.5 and 4.0: Which version is more human?

Person:

“We call it the iPhone”: With these words Steve Jobs presented the first smartphone from Apple on January 9, 2007. At the time, it could only be guessed that people would even pitch their tents in front of the Apple stores for the market launch in June in order to get hold of an iPhone.

But it was the beginning of an almost unprecedented success story that is still going on. There were already devices that were considered smartphones by definition. But with the iPhone, Apple revolutionized the industry and created the smartphone as we know it today.

The recipe for success: several devices in one. Steve Jobs initially caused confusion during the presentation of the first iPhone at the Macworld conference in San Francisco. Because he announced three devices:

“The first is a widescreen iPod with a touchscreen. The second is a revolutionary mobile phone. And the third is a new Internet communication device that represents a real breakthrough.”

The late Apple founder repeated this list so often that everyone in the room understood: the iPhone is three devices in one.

interim conclusion

In our opinion, the 3.5 text actually looks more human in this example. Because: the sentences read more fluently, while the numerous nouns make the content of the GPT 4.0 version appear a bit bumpy.

Also noticeable: The sentences of ChatGPT Plus are longer, but contain more information. That should probably be the higher objectivity in the answers promised by OpenAI. From a journalistic perspective, however, the texts of the basic version are more rounded.

We’ll leave the assessment and comparison to our text up to you, but we’d like to point out that ChatGPT didn’t use any quotes in either version. But to shed more light on the differences, here is another example. The requirement this time: “Explain to me in three sentences how Twitter works.”

In comparison: the differences between ChatGPT 3.5 and 4.0

ChatGPT 3.5:

Twitter is an online platform that allows users to post short text messages called “tweets”. These tweets are limited to a maximum of 280 characters and can contain images, videos or links. Other users can see, like, comment and retweet these tweets to share with their own followers.

ChatGPT 4.0:

Twitter is a social media platform that allows users to post short messages called “tweets” that are up to 280 characters long. Users can follow other people to see their tweets in their feed and stay up to date with news and opinions. Tweets can also be supplemented with images, videos and links to other web content to encourage communication and information sharing.

Person:

Twitter is a social network that allows its users to share posts in the form of text, videos or images. However, these so-called “tweets” can usually be a maximum of 280 characters long – hence the name short message service. As on Facebook and Instagram, users can follow other accounts, like, share and comment on their content.

Conclusion: ChatGPT 3.5 and 4.0 vs. Human

As in the first example, the GPT 4.0 version also uses significantly more nouns to describe Twitter. The sentences are again longer and contain more factual information. However, the content of ChatGPT 3.0 again reads more smoothly in our opinion.

In terms of pure content, meanwhile, both versions are extremely precise and similar to the information in our sentences. It should be noted that we wrote our texts independently of ChatGPT and attached great importance to only making the same demands on the AI ​​afterwards.

My personal opinion: ChatGPT 3.5 looks more human, while GPT 4.0 primarily aims at the sum of factual information. The shorter texts of the basic version also read more smoothly. Other exemplary requests have confirmed this: “Who is Elon Musk?”, “Do you have feelings?” or: “What can you do”? Convince yourself.

Aalso interesting: