The long-awaited upgrade to the ChatGPT AI chatbot, the GPT-5 has been released by the manufacturer Openai.
Openai's first offering, ChatGPT (one of the most anticipated releases in Silicon Valley after its GPT-3 model) kickstarted the current AI boom in late 2022.
“The GPT-3 felt like talking to high school students,” said Sam Altman, CEO of Openai.
“GPT-4, it could have been like talking to a college student, but with GPT-5, it's like talking to an expert now.
At the launch event, Openai claimed that the new chatbot, which will be released to all ChatGPT users on Thursday, is more than a simple upgrade to the previous product.
According to Openai, the new model is a standardized test used to rank models, surpassing the Chatbot competitions such as Google, X, and Antropic.
Openai claims that writing prose and advanced computer code, solving academic questions from mathematics to law, and not only assisting with medical-related questions, it is designed to be easier and more natural to be safer than its predecessors.
“It's an incredible superpower in high demand,” Altman argued.
GPT-5. Photo: Openai
The model is also more intelligent about how it uses its own brain power and therefore expensive computing resources, according to Openai.
This is a hybrid of previous chatbots, a slower, more computing-intensive “inference” model, like Openai's Deep Research.
Based on the user's request, the model determines how much “thinking” is needed before responding, rather than asking the user to switch between different models.
GPT-5. Photo: Openai
The GPT-5 “AI enthusiasts who expected to represent artificial general information [AGI]”It's going to be disappointing.
Despite this being Openai's goal, we charged a “major upgrade” to GPT-5 and a “critical step along the path to AGI” and a GPT-5.
But they are clearly not there yet.
For more accessible video players, use the Chrome browser
July: “ChatGpt is the partner I've always wanted.”
The actual test of the GPT-5 is whether to sell or not.
Openai is projected to spend $8 billion (£6 billion) on top of $5 billion (£3.7 billion) last year, and while it is expected to make a profit this year, the business case for an increasingly powerful AI model is not yet clear to many investors.
A single training run of the GPT-5 is rumoured to cost $500 million (£373 million), so the new model is expected to be extremely useful for business users.
Despite the extremely smooth demonstration of coding skills in the launch presentation, GPT-5, which is a building online language learning game in seconds, needs to prove the value of professional coding.
Many in the tech industry prefer the Claude model of human AI to write code. Openai and its investors hope that GPT-5 will change it.
AI experts test the “hastique” trends in GPT-5. This claims Openai has improved with GPT-5.
But the false or strange answer is the problem of making all large-scale generated AI models a dog.
“We've seen a lot of experience in AI commentators,” said Gary Marcus, a cognitive scientist at New York University.
“But that doesn't mean that it's a critical step towards the best path to AI that we can trust.”
