On Sunday, Sam Altman, CEO of OpenAI, provided a glimpse into the company’s forthcoming artificial intelligence model, GPT-5. In a post on social media, he shared a screenshot of a question posed to the model and its generated response, showcasing enhancements in text generation capabilities. Although there have been reports suggesting that GPT-5 could be launched in August, OpenAI has yet to confirm these claims. Altman previously indicated that this model will be the first from OpenAI to integrate both the GPT series and the O-series.
Sneak Peek of GPT-5 Before Its Release
In a recent post on X, Altman praised the AMC series “Pantheon.” When a user inquired if GPT-5 recommended the show, Altman responded by sharing a screenshot indicating that the new large language model (LLM) did, in fact, provide the recommendation.
Altman asked GPT-5, “What is the most thought-provoking show about AI?” The model replied with a carefully selected list of shows, including information on where they can be streamed or aired, accompanied by brief, spoiler-free descriptions.
To analyze the differences between the two AI models’ capabilities, we at Gadgets 360 replicated Altman’s prompt using GPT-4o, which is accessible at no cost. A preliminary examination suggests that GPT-5 offers a more conversational style compared to its predecessor.
Response from GPT-4o with the same prompt
Furthermore, GPT-4o’s output adhered to a more traditional structure, utilizing bullet points and subheadings. In contrast, GPT-5 responses exhibit a more fluid and readable format. The forthcoming model also included more detailed factual information in its recommendations, such as Rotten Tomatoes ratings and insights into viewer perceptions, while GPT-4o’s answers tended to be more generic and focused on plot summaries.
This disparity in responses may align with Altman’s earlier statements emphasizing that GPT-5 would blend the conversational style of GPT with the analytical capabilities of the O-series. Additionally, the anticipated LLM is expected to come equipped with access to all of OpenAI’s tools, including Canvas, Web Search, Image Generation, Advanced Voice Mode, and the new ChatGPT Agent.
