A significant upgrade to GPT-4 – OpenAI’s most advanced publicly available model – has made ChatGPT more conversational and concise, the startup has revealed.
A more conversational approach brings ChatGPT into line with competitors Gemini and Claude, which produce more human-like responses than OpenAI’s trailblazing creation – at least according to our tests.
But will this new update give ChatGPT the edge over its rivals? At present, that remains unclear – but as more and more users are granted access in the coming weeks, we’ll soon find out.
ChatGPT Gets Better at… Chatting
An updated version of GPT-4 Turbo – the Large Language Model (LLM) that powers the paid version of ChatGPT – has been released. According to OpenAI, it’s much more conversational and performed better than the previous GPT-4 model on several tests.
The Microsoft-backed AI startup says they’ve improved the chatbot’s capabilities “in writing, math, logical reasoning, and coding” – so just about everything, then.
Aside from that, they haven’t given much away – announcing the update through Twitter rather than an explanatory post on their blog – although the company did add that ChatGPT will be “more direct, less verbose, and use more conversational language.”
This is sorely needed, too – out of the three major chatbots that have captured the world’s attention, ChatGPT is known for sounding the least “human” in its answers.
This just in! View
the top business tech deals for 2024 👨💻
Claude Sonnet and Opus – the two most advanced language models created by Anthropic – tend to respond with much more emotion compared to ChatGPT.
Who Can Access the New Version of GPT-4 Turbo?
Unfortunately, the new version of GPT-4 Turbo is only available to paying ChatGPT customers at the moment. This includes ChatGPT Plus users and businesses using the enterprise API.
However, if you’re a free user of ChatGPT, you’ll probably have to wait until GPT-5 is released to get your hands on GPT-4. Currently, GPT-3.5 – a less sophisticated LLM from the same family – powers the free version of ChatGPT.
In fact, OpenAI CEO Sam Altman recently said himself that GPT-4 “sucks” – which you’d hope signals something much better isn’t far around the corner.
How Does GPT-4 Compare to Other Language Models?
There’s no doubt that GPT-4 is a highly capable language model, but according to Anthropic, Claude 3 Opus – its most advanced language model – comes up trumps on several benchmark tests.
It’s hard to split ChatGPT and Gemini, as they both perform well on different tasks – but Gemini was the one that edged it when we tested both of their free and paid services. However, this GPT-4 upgrade could easily flip the results – after all, the margins were fine.
Whatever chatbot you’re using at work, make sure to follow your company’s AI guidelines, be transparent about your usage, and check what these AI companies can do with the data users input into their products.
