OpenAI Files GPT-5 Trademark Application
OpenAI has applied to trademark GPT-5, the next iteration of its large language model. The United States Patent and Trademark Office filing submitted on July 18 suggests the LLM is in the works, although GPT-4 is only a few months old at this point.
GPT-5 TM
The technical details of GPT-5 are only speculative, but it could be an enormous improvement on the current iteration. GPT-4 marked a significant improvement in nearly every metric compared to GPT-3, which had a similar improvement over GPT-2. The trademark submission keeps it simple. GPT-5 is “downloadable computer software and programs” for activities like “the artificial production of human speech and text,” for “natural language processing, generation, understanding, and analysis,” and various other tasks like translation and transcription. The trademark also goes for the very broad description of generative AI, with GPT -5’s trademark claiming it is capable of “developing, running and analyzing algorithms that are able to learn to analyze, classify, and take actions in response to exposure to data,” and that it can be used for “developing and implementing artificial neural networks.”
There’s no date for when GPT-5 might debut in the submission, though some rumors suggest it might be done training by 2024. That said, OpenAI CEO Sam Altman has publicly claimed that GPT-5 is not being trained right now. That doesn’t mean other aspects of the LLM aren’t under development, of course. It might also just be an early bureaucratic checkmark for the eventual creation of GPT-5 that OpenAI wanted to get out of the way, like buying a domain name for a product that hasn’t been finished.
It’s not as though OpenAI doesn’t have plenty of other projects to keep it busy. That includes the long-rumored G3PO open-source LLM, an entirely different track for its LLM research but one that may now be accelerating after Meta introduced the open-source Llama 2 LLM. OpenAI also recently started some journalism-based AI projects and unveiled a Custom Instructions feature that may presage a fully personalized ChatGPT assistant. That’s on top of a supposed app store, where developers can shop for different ways to leverage OpenAI’s models, set to roll out in the near future.
Follow @voicebotai Follow @erichschwartz
OpenAI Personalizes ChatGPT With New Custom Instructions Feature
OpenAI Debuts GPT-4 Multi-Modal Generative AI Model With a Sense of Humor
OpenAI Steps Up Work on ‘G3PO’ Open-Source LLM to Compete with Meta: Report