OpenAI has filed a trademark application for “GPT-5,” providing a glimpse of a potential successor to its successful language model, GPT-4.
The application, spotted on July 18 by Windows Latest, describes “downloadable computer software for using a language model.”
Other features mentioned include “the artificial production of human speech and text” and “natural language processing, generation, understanding, and analysis,” aligning with previous trademark applications for GPT-3.5 and GPT-4.
While specific details about GPT-5 remain undisclosed, it’s likely that OpenAI will first focus on improving GPT-4 instead of immediately launching the new iteration.
GPT-4, released in March and made available via API, has gained popularity through applications like ChatGPT and integration with Microsoft’s Bing browser. However, it has been noted for occasionally producing inaccurate information and incurring high operational costs.
The company may aim to address these issues with GPT-5, but any official launch date or further information remains unknown.
Until then, the AI community eagerly awaits updates on OpenAI’s progress toward its new language model, GPT-5.