OpenAI has launched GPT-5, the latest upgrade to its ChatGPT AI chatbot. The new model goes live for all ChatGPT users Thursday.
OpenAI claims GPT-5 is a major leap beyond GPT-4. According to CEO Sam Altman, GPT-3 felt “like talking to a high school student,” GPT-4 like a college student, and GPT-5 now “like talking to an expert, a PhD-level expert in anything.”
The model reportedly beats competitors like Google and Anthropic on standard benchmarks. It’s better at natural communication, writing prose, advanced coding, solving academic problems from math to law, and answering healthcare questions. Safety improvements are also a focus. 
“It’s an incredible superpower on demand,” Sam Altman said at the launch.
GPT-5 uses computing resources more intelligently. It combines fast chatbot responses with slower, deeper reasoning from models like OpenAI’s Deep Research. The AI decides how much “thinking” it needs per request—no model switching required.
Despite hype, GPT-5 isn’t artificial general intelligence (AGI). Altman called it a “significant step along the path to AGI,” but the AI community isn’t convinced.
OpenAI spent a rumored $500 million training GPT-5, on top of billions already invested. It has to prove its value, especially for business users and coders. The launch presentation highlighted coding skills, but many still prefer Anthropic’s Claude for programming.
Hallucinations—wrong or bizarre answers—remain a concern, despite OpenAI’s claims of improvement.
Gary Marcus, NYU cognitive scientist, commented:
“Shiny things are always fun to play with, and I fully expect GPT-5 to be the shiniest so far,”
“But that doesn’t mean that it is a critical step on the optimal path to AI that we can trust.”
GPT-5’s commercial success is still uncertain as OpenAI aims to turn a profit amid huge costs.
