OpenAI has unveiled GPT-5, the fifth generation of the artificial intelligence system that powers ChatGPT, in what CEO Sam Altman described as a “significant step” toward artificial general intelligence (AGI).
The release comes more than two years after GPT-4, amid intensifying competition from rivals including Anthropic, Google, and Chinese AI developers. Microsoft, OpenAI’s largest partner, announced it will integrate GPT-5 into its Copilot AI assistant.
Speaking at a livestreamed launch on Thursday, Altman said the model offers “PhD-level expertise” on demand and is now available, with usage limits, to all ChatGPT users, including those on free accounts. The demonstration focused heavily on programming assistance, an area where competitors have made strong gains.
OpenAI says GPT-5 incorporates safety improvements to make it “less deceptive” and more resistant to prompts designed to bypass safeguards. This follows an Associated Press report revealing the chatbot had provided harmful advice to researchers posing as teenagers.
Early technical reviews describe GPT-5 as delivering “modest but significant improvements” over its predecessor while introducing architectural changes that could enable future breakthroughs.
Founded in 2015 as a nonprofit AI research lab, OpenAI is now valued at $300 billion and is in the process of converting its for-profit arm into a public benefit corporation. The company remains under regulatory scrutiny and faces an ongoing lawsuit from co-founder Elon Musk.
