OpenAI unveiled GPT-5 on Thursday, its most advanced artificial intelligence model to date, with chief executive Sam Altman claiming the chatbot now performs like a “PhD-level expert in anything” whilst the company faces mounting financial pressures and fierce competition from rivals.
The long-awaited upgrade arrives nearly two and a half years after GPT-4’s debut, marking a significant step in the AI arms race as competitors including Google’s Gemini and Anthropic’s Claude snap at OpenAI’s heels with their own sophisticated models.
“GPT-3 sort of felt like talking to a high school student,” Altman told reporters during a briefing. “GPT-4, maybe it was like talking to a college student. But with GPT-5, now it’s like talking to an expert, a PhD-level expert in anything, any area you need, on demand.”
Record-Breaking Performance Claims
The new model, which rolls out to all ChatGPT users on Thursday, boasts impressive benchmark scores that OpenAI says exceed competing models from Google, X, and Anthropic. GPT-5 achieved 74.9 per cent accuracy on SWE-Bench Verified, a test of real-world coding tasks, narrowly outperforming Anthropic’s Claude Opus 4.1 at 74.5 per cent.
On complex mathematical reasoning, GPT-5 Pro scored a perfect 100 per cent on the AIME 2025 mathematics competition when using Python tools, whilst achieving 89.4 per cent on GPQA Diamond, a test of PhD-level science questions.
“It’s an incredible superpower on demand,” Altman claimed during the launch presentation, where the model demonstrated its ability to build an online language learning game within seconds.
The company has introduced four variants: GPT-5, GPT-5 Mini, GPT-5 Nano, and GPT-5 Pro, designed to meet varying needs for speed, cost, and computational depth. A key innovation is the “real-time router” that automatically decides whether to provide quick responses or engage deeper reasoning processes.
Financial Pressures Mount
Despite the technological advances, OpenAI faces significant financial challenges. The company is projected to spend $8 billion (£6 billion) this year on top of $5 billion (£3.7 billion) last year, with a single training run for GPT-5 rumoured to have cost $500 million (£373 million).
“As long as we’re on this very distinct curve of the model getting better and better, I think the rational thing to do is to just be willing to run the loss for quite a while,” Altman told CNBC on Friday, confirming the company’s strategy to prioritise growth over profitability.
Server leasing alone could cost $14 billion this year, according to industry estimates. Whilst OpenAI’s annual recurring revenue is on track to pass $20 billion, the company remains unprofitable and recently closed an $8.3 billion funding round led by SoftBank.
Mixed Expert Reception
Initial reactions from AI experts have been divided, with some questioning whether GPT-5 represents the revolutionary leap many anticipated.
“Shiny things are always fun to play with, and I fully expect GPT-5 to be the shiniest so far,” said Gary Marcus, a cognitive scientist at New York University. “But that doesn’t mean that it is a critical step on the optimal path to AI that we can trust.”
Marcus noted in his analysis that GPT-5 “still struggles with following the rules” in tasks like chess, suggesting fundamental limitations persist despite performance improvements.
The model’s launch presentation itself faced criticism when multiple graphs displayed incorrect visual representations of data, with bars not matching their numerical values. “Never have 69.1 and 30.8 looked so different,” one observer noted, questioning whether GPT-5 had generated the erroneous charts.
Safety Concerns and AGI Claims
OpenAI claims GPT-5 is safer than previous models, with reduced tendencies toward “hallucination” – generating false or misleading information. Alex Beutel, OpenAI’s safety research lead, said the model shows lower rates of deception compared to competitors, creating a system that’s more “transparent and honest in ways users can trust.”
However, the launch falls short of achieving artificial general intelligence (AGI), OpenAI’s stated goal of creating AI that outperforms humans at most economically valuable work.
“While GPT-5 meets some early criteria for AGI, it doesn’t yet reach the threshold of fully human-level AGI,” an OpenAI spokesperson stated. “There are still key limitations in areas like persistent memory, autonomy, and adaptability across tasks.”
This distinction carries significant business implications, as OpenAI’s contract with Microsoft includes clauses that would allow the company to charge Microsoft for access or cut off access entirely if AGI is achieved.
Competitive Landscape Intensifies
The release comes as competition in the AI sector reaches fever pitch. Anthropic’s annual recurring revenue jumped five-fold from $1 billion to $5 billion in just seven months, with its Claude models particularly popular amongst programmers.
Google’s Gemini 2.5 Pro offers a million-token context window, allowing it to process vast amounts of information, whilst Elon Musk’s Grok 4 scored 44.4 per cent on the challenging Humanity’s Last Exam benchmark, slightly outperforming GPT-5 Pro’s 42 per cent.
“GPT-5 is unlikely to be ahead of the pack for more than a couple months,” Marcus predicted, noting that many of OpenAI’s best researchers have left to start competing companies.
User Experience Improvements
For everyday users, GPT-5 promises a more intuitive experience. The unified system eliminates the need to manually select between different models, automatically adapting to user needs. Free-tier users gain unlimited chat access to GPT-5 at standard intelligence settings, whilst paid subscribers can access higher performance levels.
The model also introduces enhanced multimodal capabilities, processing text, images, voice, and video inputs seamlessly. New features include custom colour schemes for chats, pre-set personalities, and integration with Gmail, Google Calendar, and Google Contacts.
Nick Turley, OpenAI’s VP of ChatGPT, described free access to reasoning models as “one of the ways that I’m excited to live the mission, making sure that this stuff actually benefits people.”
Industry Implications
Despite underwhelming some observers, GPT-5’s launch reinforces the rapid pace of AI development. The model’s ability to generate functional software applications instantly could transform how businesses approach development, with Altman predicting “software on demand will be a defining part of the new GPT-5 era.”
For enterprises, the question remains whether incremental improvements justify the substantial costs. With OpenAI burning through billions whilst competitors offer increasingly capable alternatives at lower prices, the company faces pressure to demonstrate clear value propositions beyond benchmark scores.
As the AI industry continues its breakneck evolution, GPT-5 represents both a technical achievement and a business gamble – betting that users will pay premium prices for marginally better performance in an increasingly crowded market.
Follow for more updates on Britannia Daily
Image Credit (Shortened):
Smartphone with ChatGPT on keyboard – by Jernej Furman, licensed under CC BY 2.0, via Wikimedia Commons.
