OpenAI, under Sam Altman's leadership, has reportedly applied for trademarks for "GPT-6" and "GPT-7" in China, signaling the company's continued efforts in developing large language models (LLMs).
According to records from the Trademark Office of the China National Intellectual Property Administration, OpenAI OpCo, the company's entity in China, submitted two trademark applications for GPT-6 and two more filings for GPT-7. Notably, OpenAI's services are not currently accessible in China or Hong Kong.
OpenAI refrained from commenting on the reported trademark applications. Since the introduction of ChatGPT a year ago, the organization has been pushing the boundaries of its deep learning LLMs.
ChatGPT initially stemmed from GPT-3.5, boasting 175 billion parameters. Altman, the mind behind AI chatbot ChatGPT and currently back at the helm after a tumultuous CEO transition, clarified that the company isn't presently training GPT-5, the hypothetical successor to GPT-4.
Recent milestones include ChatGPT achieving 100 million weekly active users, accompanied by the launch of the new GPT-4 Turbo model. This upgraded version offers enhanced capabilities, cost-effectiveness, and supports a 128K context window.
During the company's inaugural developer conference, Altman highlighted that over two million developers leverage ChatGPT, with more than 92 percent of Fortune 500 companies utilizing the platform.
GPT-4 Turbo's 128k context window allows it to accommodate the equivalent of over 300 pages of text in a single prompt. OpenAI optimized its performance, resulting in GPT-4 Turbo being offered at three times cheaper for input tokens and twice cheaper for output tokens compared to GPT-4.
Beyond GPT-4 Turbo, OpenAI is rolling out a new iteration of GPT-3.5 Turbo, which inherently supports a 16K context window.
Additionally, the company introduced the Assistants API, empowering developers to create agent-like experiences within their own applications.
(With Agency Inputs)