what we know about the next OpenAI chatbot

OpenAI is reportedly preparing to release a more powerful version of ChatGPT in the coming months.

According to two sources who spoke to Business Insider, the new artificial intelligence model, known as GPT-5, is expected to arrive this summer. Before the launch, some companies reportedly tried out a demo version of the tool to test its improved capabilities.

This technology is part of OpenAI’s futuristic search for artificial general intelligence (AGI), i.e. systems smarter than humans.

This lofty science fiction premise predicts an artificial intelligence that can think for itself, thus creating more similar models of artificial intelligence without the need for human supervision. Depending on who you ask, such a breakthrough could either destroy the world or supercharge it.

As the race to build the best AI heats up, here’s everything you need to know about GPT-5.

What is GPT-5?

GPT-5 is the sequel to GPT-4, OpenAI’s fourth-generation chatbot, which requires a monthly fee to use.

GPT stands for Pre-Trained Generative Transformer, which is an AI engine built and refined by OpenAI to power various versions of ChatGPT. Just like the processor in your computer, each new edition of the chatbot runs on a completely new GPT with greater capabilities.

With GPT-4, the AI ​​chatbot can provide human-like responses and even recognize and generate images and speech. Its successor, GPT-5, is said to offer better personalization, make fewer errors and support more types of content, including eventually video.

Others, such as Google and Meta, have released their own GPTs with their own names, all of which are known collectively as large language models.

Human brain-based AI systems have the ability to generate text within a conversation.

Is GPT-5 trained?

The latest report shows that OpenAI has started GPT-5 training in preparation for the AI ​​model’s launch in the middle of this year. According to Business Insider, after completing the training, the system will undergo multiple stages of security testing.

As part of this process, the bot will be combined into a “red team”. This is a technique in which internal and external testers examine it and provide feedback on its strengths and weaknesses.

The report follows speculation that GPT-5’s learning process may have recently begun, based on a recent tweet from an OpenAI official.

In January, one of the tech company’s top researchers suggested that OpenAI was exercising a much larger GPU than usual. The report followed a separate tweet from OpenAI’s co-founder and CEO that detailed how the company has expanded its computing resources.

In short, a graphics processing unit. The graphics processor works like a calculator that helps the AI ​​model establish connections between different types of data, such as linking an image to its corresponding text description.

For his part, Altman confirmed that his company worked on GPT-5 at least twice last fall.

According to two people who attended the event, the first occurred during a conversation at the alumni reunion of his former venture capital firm Y Combinator last September. Altman said GPT-5 and its successor, GPT-6, “were in the bag” and were better than their predecessors.

In November, he went public with its existence, telling the Financial Times that OpenAI was working on GPT-5, although he did not reveal its release date.

Recently, a report claimed that the head of OpenAI came up with a bold plan to acquire the massive amounts of GPUs needed to train larger artificial intelligence models.

To overcome supply shortages hindering technological innovation, Altman wants to raise as much as $7 trillion for rapid chip production with the help of a global network of investors, governments and energy suppliers, according to the Wall Street Journal.

What can GPT-5 do?

In February, the head of OpenAI spoke about GPT-5 at the World Government Summit in Dubai. He said that the next version of ChatGPT will be smarter than its predecessors.

“It’s a bigger deal than it seems, because what makes these models so magical is their versatility,” Altman explained. “So if he’s a little smarter, he’s a little better at everything.”

Mentioning his brain’s capabilities, Altman told the FT that GPT-5 would require more data to train. The plan was to use publicly available data sets from the Internet and large-scale, proprietary data sets from organizations. The last of these would involve long-form writing or conversations in any format.

In mid-January, on Bill Gates’ Unconfuse Me podcast, Altman said: “Multimodality is definitely going to be important. This means speech in and out. Pictures. Finally a video. It’s obvious that people really want this. We will be able to go much further, but perhaps the most important areas of progress will be in reasoning skills.”

He continued: “Right now, GPT-4 can only reason in a very limited way. Also reliability. If you ask GPT-4 most questions 10,000 times, one of those 10,000 times will probably be quite good, but it doesn’t always know which, and you want to get the best 10,000 answer every time, so the increase in reliability will be important.”

GPT-5 vs. GPT-4

So how can it beat GPT-4? First of all, it will have to outdo GPT-4 Turbo, the next-generation model that OpenAI released in November for paying subscribers.

The company’s most advanced AI chatbot has knowledge of world events up to April 2023 compared to 2021 for GPT-4; can analyze even longer prompts of up to 128,000 tokens, or about the length of a 300-page book; it is better to follow the instructions; and can automatically switch between tools, including the Dall-E 3 image generator and Bing search engine, based on user requests.

Both OpenAI and several researchers have also tested the chatbot on real-life exams. GPT-4 has been shown to have a decent chance of passing the challenging Certified Financial Analyst (CFA) exam. He scored in the 90th percentile on the bar exam, passed the SAT reading and writing section and was ranked 99th-100th. percentile on the 2020 USA Biology Olympiad semifinal exam.

When will GPT-5 be available?

At the time of writing, OpenAI has not announced a release date for GPT-5.

It’s also unclear whether this had an impact on the turmoil at OpenAI late last year. On November 17, Mr. Altman was dismissed by the company’s management board. After five days of confusion that was symptomatic of contentious viewpoints about the future of artificial intelligence, Altman returned to the helm with a new board of directors.

Interestingly, some ChatGPT users previously claimed that the bot told them it was running on a new AI model called GPT-4.5 Turbo, but this turned out to be an error.

If OpenAI’s GPT release schedule tells us anything, it’s that the interval between updates is getting shorter. GPT-1 arrived in June 2018, followed by GPT-2 in February 2019, then GPT-3 in June 2020, and the current free version of ChatGPT (GPT 3.5) arrived in December 2022, and GPT -4 arrived just three months later in March 2023. Recent months have also seen more frequent updates, including a “turbo” version of the bot.

So can June be a magical month again? Or will OpenAI wait until its Developer Day conference, likely in November, to announce GPT-5? We’ll have to wait and see.