OpenAI Announces its Newest Model GPT-4 Turbo
This week, OpenAI recently announced multiple new features for ChatGPT through a blog on its official website. They also released new information on other AI-powered tools during its recent developer conference.
It seems that 2023 will be a year of GPT launches as back in March, OpenAI launched GPT-4 and is expected to release GPT Turbo later this year. Some of the new features of GPT-4 Turbo will be a welcomed addition for users, particularly its new knowledge cutoff.
According to OpenAI, the new cutoff data will move from September 2021 to April 2023. This will give users a greater current context for prompts. But that’s not all. CEO Sam Altman also expressed the desire to bring their GPT models to the point where the current information cutoff is no longer an issue.
However, how that would be possible is still unknown. This is one of OpenAI’s major headaches as the company is facing lawsuits and threats of lawsuits from multiple groups due to how information was allegedly used to train models without the creator’s consent.
For developers who are keen on using ChatGPT’s API, Altman hinted that prices will be going down. In the same blog post, it provided two examples of this. First with DALL-E 3’s pricing which started at $0.04 per image generated while also stating that GPT-4 Turbo with vision would see a cost of $0.00765 per image.
Now for those who are becoming expert prompt engineers, GPT-4 Turbo will support up to 128,000 tokens of context. Allowing users to create extremely long and detailed prompts. As we saw this past March, when users are allowed to build off prompts of greater size, what is generated tends to be quite incredible.
Another upgrade will be in how the model will be able to follow user-inputted instructions. By keeping itself consistent with what is generated, this feature would likely benefit developers who are using the chatbot to help them generate code.
Finally, GPT’s user interface will be smarter understand your request, and pick the correct tool. So instead of the familiar dropdown menu where users had to pick the tools they wanted to use, the AI would now pick the tools based on your input.
So if you’re requesting an image, DALL-E 3 will respond to your prompt and generate the appropriate image. These are some interesting features coming from OpenAI and how their competitors, practically Google, will respond will be key in shaping the current AI contest. CNET provided a great first impressions video of OpenAI’s announcement below:
Originally posted on OpenDataScience.com
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Interested in attending an ODSC event? Learn more about our upcoming events here.
