OpenAI has officially launched GPT-5, the latest iteration of its flagship large language model that powers ChatGPT, with CEO Sam Altman describing it as a “major upgrade” and a crucial step toward achieving artificial general intelligence (AGI).
The $300 billion company announced on Thursday that GPT-5 sets a new standard in areas such as mathematics, science, and coding. The model will be accessible to all ChatGPT users, although free-tier users will face some usage limits.
GPT-5 arrives nearly two and a half years after the debut of GPT-4, which had already marked a significant leap in AI capabilities. Since then, anticipation has built among investors and users eager to see OpenAI’s next breakthrough.
Altman highlighted GPT-5’s strength in what he called “vibe coding,” a feature allowing users to instruct the AI via text prompts to create new software on demand—a development he suggested will define the GPT-5 era.
Since ChatGPT’s launch almost three years ago, OpenAI has surged in popularity, now boasting more than 700 million weekly active users. The company is reportedly exploring a valuation near $500 billion, which would position it as the world’s most valuable private tech firm.
Backers hope GPT-5’s performance will justify their confidence that OpenAI could one day become a multi-trillion-dollar enterprise. However, initial reactions have been mixed.
“GPT-5 meets expectations in technical performance, exceeds in task reasoning and coding, and underwhelms in [other areas],” said Gartner analyst Chirag Dekate, adding that it “does not appear to cross the AGI threshold” and shows only incremental improvements in writing quality.
Still, even modest gains are significant in the competitive AI landscape. Coding remains one of the most commercially valuable AI applications. OpenAI’s GPT-5 marginally outperformed Anthropic’s latest model on SWE-bench Verified, a widely recognized coding benchmark.
This edge has attracted key users such as Anysphere, a $10 billion start-up behind Cursor, a popular AI-powered code assistant. CEO Michael Truell described GPT-5 as “remarkably intelligent.” Cursor, previously an Anthropic customer, shifting to GPT-5 could substantially boost OpenAI’s annual recurring revenue (ARR), currently estimated at $12 billion and projected to reach $20 billion by 2025.
Despite GPT-5’s advancements, Elon Musk’s xAI recently launched Grok 4 Heavy, which still outperforms GPT-5 on some reasoning and knowledge tests.
In trials, GPT-5 demonstrated greater accuracy and was less prone to “hallucinations”—instances where the AI fabricates information. However, it is also the first flagship model labeled as “high risk” for potential misuse, including the creation of biological weapons or novel viruses.
OpenAI stated, “While we do not have definitive evidence that this model could meaningfully help novices to create severe biological harm — our defined threshold for high capability — we have chosen to take a precautionary approach.”
OpenAI primarily generates revenue from subscription plans, ranging from $20 to $200 per month for individuals, with custom pricing available for businesses and educational institutions. This week, the company also released its first “open” AI model since ChatGPT’s launch, aiming to counter rising competition from Chinese start-up DeepSeek and others.
Nick Turley, OpenAI’s head of product for ChatGPT, summed up early impressions: “The vibes of this model are really good. But at the end of the day, I think people are just going to have to feel it.”
