OpenAI has announced the launch of GPT-5, describing it as its “best AI system yet” and a significant leap over previous models. The company says the system offers state-of-the-art performance in areas including coding, mathematics, writing, health, and visual perception.

The model is available to all ChatGPT users, with Plus subscribers receiving higher usage limits and Pro subscribers gaining access to GPT-5 Pro, a version with extended reasoning capabilities for more comprehensive responses.

According to OpenAI, GPT-5 is a unified system consisting of a standard model for most queries, a “thinking” model for more complex problems, and a real-time router that determines which to use based on conversation type, complexity, tool needs, and user intent. The router is trained on usage signals, such as when users switch models or express a preference for a certain style of answer.

Once usage limits are reached, a smaller version of each model handles remaining queries. OpenAI says it plans to integrate these capabilities into a single model in future.

Performance and Capabilities

The company says GPT-5 delivers faster and more accurate responses than earlier versions, while reducing hallucinations, improving instruction-following, and minimising sycophancy. It highlights improvements in three core areas: writing, coding, and health.

In coding, OpenAI claims GPT-5 shows notable gains in complex front-end generation and debugging larger repositories. The system is said to produce responsive websites, apps, and games from a single prompt, with improved understanding of design elements such as spacing, typography, and white space.

For writing, the model is described as OpenAI’s “most capable collaborator yet”, with an ability to sustain complex literary forms and handle structural ambiguity. It can assist with tasks ranging from creative writing to drafting and editing professional documents.

In health-related queries, GPT-5 reportedly scores higher than any previous model on HealthBench, an evaluation based on realistic scenarios and physician-defined criteria. OpenAI says it can flag potential concerns and ask clarifying questions to provide more helpful answers, while adapting to a user’s context and geography. The company stressed that ChatGPT is not a replacement for medical professionals, describing it instead as a partner to help users interpret information and prepare for consultations.

Industry Reaction and Competition

Speaking ahead of the launch, OpenAI chief executive Sam Altman said GPT-5 represented a new stage in ChatGPT’s development. “I think having something like GPT-5 would be pretty much unimaginable at any previous time in human history,” he said.

Altman compared the new model to a human expert. “GPT-3 sort of felt to me like talking to a high school student… 4 felt like you’re kind of talking to a college student. GPT-5 is the first time that it really feels like talking to an expert in any topic, like a PhD-level expert.”

The launch comes amid a competitive race among major technology firms to develop the most advanced AI chatbots. Elon Musk recently claimed his own system, Grok, was “better than PhD level in everything” and “the world’s smartest AI”.

Recommended reading

• ChatGPT Search Inches Toward Strict EU Oversight as Users Surge
• Is ChatGPT Making You Dumb? Possibly, MIT Says
• ‘Sycophant-y’ ChatGPT Rollback Sparks AI Personality Debate

Not all observers share the same level of enthusiasm. Prof Carissa Véliz of the Institute for Ethics in AI suggested the significance of the release may be more limited than OpenAI’s marketing suggests. “These systems, as impressive as they are, haven’t been able to be really profitable,” she said, adding that they can only mimic rather than truly emulate human reasoning. “There is a fear that we need to keep up the hype, or else the bubble might burst, and so it might be that it’s mostly marketing.”

Tensions With Anthropic

The launch also follows a dispute between OpenAI and rival Anthropic. Anthropic recently revoked OpenAI’s access to its application programming interface (API), alleging the company had violated its terms of service by using its coding tools ahead of GPT-5’s launch.

An OpenAI spokesperson described it as “industry standard” to evaluate other AI systems when assessing progress and safety. “While we respect Anthropic’s decision to cut off our API access, it’s disappointing considering our API remains available to them,” the spokesperson said.

In a blog post accompanying the launch, OpenAI said GPT-5 was designed to feel more responsive and personal than earlier technologies, particularly for vulnerable users. However, it added that the system would not give definitive answers to personal questions such as “Should I break up with my boyfriend?”, instead helping users “think it through” by asking questions and weighing options.

UK Think Tank Warns That AI Regulation is Lagging

In response to the launch, the Ada Lovelace Institute has warned that GPT-5 “reinforces the growing gap between AI capabilities, and our ability to govern them.”

The UK-based research group said the “PhD-level” intelligence claimed for GPT-5 highlights the urgent need for regulation, with 72% of the UK public saying laws would increase their comfort with AI.

“Almost three years since the Bletchley AI Summit, the only actors making the decisions regarding whether such systems are safe enough to release are the companies themselves,” the institute said, warning governments currently lack the power to halt unsafe AI systems.