OpenAI has released ChatGPT-5 for free to all users, showcasing major improvements in speed, intelligence, and task handling. CEO Sam Altman says it performs like a PhD-level expert and marks a step toward AGI. The model emphasizes safety and honesty, while also powering U.S. government tools and new open-source AI models.
In a landmark move signaling a new era in artificial intelligence, OpenAI announced the release of ChatGPT-5 on Thursday, making the cutting-edge language model available for free to all users worldwide. The company heralded the release as a major advancement in AI capabilities, boasting performance that surpasses previous versions across a broad range of tasks.
OpenAI’s co-founder and CEO, Sam Altman, described ChatGPT-5 as a dramatic improvement over earlier iterations, positioning it as the most capable and intelligent model the organization has developed to date.
“GPT-3 felt to me like talking to a high school student — ask a question, maybe you get a right answer, maybe you'll get something crazy,” Altman said. “GPT-4 felt like a college student. GPT-5 is the first time it really feels like talking to a PhD-level expert in any topic.”
According to OpenAI, ChatGPT-5’s deployment comes at a time when nearly 700 million people use ChatGPT on a weekly basis. The model is designed to perform a wide array of tasks — from generating complex code and writing articles to supporting healthcare and scientific research — with enhanced speed, reliability, and contextual understanding.
Table of contents [Show]
Altman noted that while GPT-5 represents a “significant step” toward achieving artificial general intelligence (AGI), it still lacks one key element: continuous learning. The model does not learn from interactions once deployed, a feature Altman believes is essential for true AGI.
“This is not a model that continuously learns as it is deployed from new things it finds, which is something that, to me, feels like it should be part of an AGI,” he said. “But the level of capability here is a huge improvement.”
The company emphasized that GPT-5 excels in acting autonomously as a digital assistant, capable of managing complex tasks independently. Michelle Pokrass, a key member of the development team, explained that GPT-5 functions more effectively as an “AI agent,” making it suitable for applications like workflow automation, software development, and customer support.
One of the standout features of ChatGPT-5 is its ability to create software programs on demand — a capability Altman referred to as “vibe coding.” During the live demonstration, the model was asked to create an app for learning the French language. Within moments, the AI generated a functional prototype, highlighting its potential to transform software development across industries.
OpenAI asserts that GPT-5 currently leads the industry in multiple verticals including healthcare, education, and creative writing, setting a new bar for rival firms such as Google and Microsoft, which have invested heavily in competing AI technologies.
Amid growing global concerns about the misuse of AI, OpenAI has placed a strong emphasis on model safety and trustworthiness. Alex Beutel, who leads safety research at OpenAI, said the model was trained with specific focus on minimizing deception and avoiding harmful content.
“We built evaluations to measure the prevalence of deception and trained the model to be honest,” Beutel stated. “GPT-5 is designed to generate safe completions and provide helpful responses without enabling malicious use.”
In a related development, OpenAI also announced a partnership with the U.S. government, offering ChatGPT Enterprise to executive branch workers for just $1 per year under a special agreement with the General Services Administration. The deal is aimed at helping federal agencies adopt AI tools responsibly and cost-effectively.
The company also released two open-weight models — gpt-oss-120b and gpt-oss-20b — which are available for public download and modification. These open-source models are part of OpenAI’s effort to promote transparency and foster innovation in the broader AI research community.
While the company started as a nonprofit dedicated to open research, its transformation into a capped-profit firm has attracted scrutiny. The release of these models is seen by many as a strategic response to calls for greater openness in the rapidly evolving AI landscape.
Altman suggested that even more powerful models are on the horizon. He indicated that future breakthroughs will require massive investments in computing power, but expressed confidence in OpenAI’s long-term vision.
“Obviously... you have to invest in compute at an eye-watering rate to get that, but we intend to keep doing it,” he said.
With ChatGPT-5 now freely accessible, OpenAI is reaffirming its role as a global leader in artificial intelligence, and pushing the boundaries of what machines can do in partnership with human creativity and oversight.
