OpenAI has launched its latest update today, with the new GPT-5 now available to everyone – including its free users.
With the launch of this update, it will mark the first time that users on the ChatGPT ‘Free’ plan will have access to a reasoning model that carries out an internal chain of thought before responding. Any Free user reaching a usage cap will have access restricted to use of GPT-5 Mini.
What can GPT-5 do?
The company claims that GPT-5 is smarter, faster, and a lot more useful than GPT-5, pushing benefits in creative writing, coding, and health care on its website. It also reports that the GPT-5 hallucination rate is lower, which means it will deliver responses with fewer factual errors.
OpenAI has said that the technology is writing less-robotically. In a demonstration, it showed GPT-5 writing hundreds of lines of code in seconds. It reports that the upgrade will be better at flagging potential concerns with health-related questions.
In regards to written creative expression, OpenAI said that it has its “most capable writing collaborator yet.” It claims it will translate your rough ideas into resonant writing that has literary depth and rhythm.
To showcase the difference between GPT-4o and GPT-5, OpenAI offered the prompt: “write a short poem that makes an emotional impact. A widow in Kyoto keeps finding her late husband’s socks in strange places.”
The results:
If you’re a bit torn on which is better, OpenAI explained that: “GPT‑5’s response lands the larger emotional arc with a stronger ending, clear imagery, and striking metaphors (“black flags of a country that no longer exists,” “Kyoto’s bell rolls evening down the hill”) that establish a vivid sense of culture and place. GPT‑4o’s version follows a more predictable structure and rhyme scheme, telling instead of showing (“she weeps and doesn’t tell”).”
Should humans be worried today that they’re out of a job?
You will still be required to punch in on Monday. With GPT-5, OpenAI is still a way off from Artificial General Intelligence (AGI), however CEO Sam Altman has labelled the launch as a “significant step on the path to AGI.”
OpenAI views AGI as reaching the stage where an autonomous system can outperform humans at most economically valuable work. It’s the “humans can spend their days drinking Coronas on the beach while wondering how they can afford their next Corona” benchmark.
During a call with reporters, Altman said “This is clearly a model that is generally intelligent, although I think in the way that most of us define AGI, we’re still missing something quite important, or many things quite important.
“This is not a model that continuously learns as it’s deployed from the new things it finds, which is something that to me feels like AGI. But the level of intelligence here, the level of capability, it feels like a huge improvement.”
An independent test of reasoning skills by ARC Prize found that GPT-5 achieved an ARC-AGI-2 benchmark of 9.9% at a cost of US$0.73 per task. Performing better than GPT-5 was the Elon Musk-owned Grok. In ARC Prize’s test, it found Grok 4 (Thinking) achieved 16%, but the downside is the cost of each task was priced at between $2-4.
AI use is scaling at a dramatic pace
Earlier this week OpenAI reported that ChatGPT is being used by 700 million weekly active users. This is up from 500 million reported in March of this year. It also reported five million paying business users on ChatGPT, up from three million in June.
ChatGPT’s biggest AI search competitor is Alphabet’s Google AI Overviews, which it reports has two billion monthly users across 200 countries. Alphabet CEO Sundar Pichai reported that the Google AI chatbot Gemini has more than 450 million active users.
OpenAI partner Microsoft has announced today that GPT-5 will be incorporated into its products from today, citing integration in Microsoft 365 Copilot and Microsoft Copilot for enterprise and consumer users.
