Mark Zuckerberg's Meta Plans To Take On OpenAI With GPT-4 Challenger: Report

Meta Platforms Inc. (NASDAQ:META), under the leadership of Mark Zuckerberg, is reportedly working on a new artificial intelligence system — a GPT-4 challenger to compete with Sam Altman-led OpenAI.

What Happened: Llama 2, Meta’s open-source AI language model released in July, marked the company’s initial foray into the AI landscape, distributed through Microsoft Corporation’s (NASDAQ:MSFT) Azure services to rival OpenAI’s ChatGPT and Alphabet Inc.’s (NASDAQ:GOOG) (NASDAQ:GOOGL) Google Bard. 

The forthcoming system aims to surpass Llama 2 in both capacity and capability, aiming to provide companies with advanced text generation and analysis tools, reported The Wall Street Journal, citing people familiar with the matter. 

See Also: From Sky-High To Rocky Ground: How Meta’s 11% Stock Plunge And Threads User Exodus Paves The Way For Their Bold New Web Launch

According to the report, the Facebook parent plans to train its GPT-4 challenger in early 2024.

While Meta collaborated with Microsoft to offer Llama 2 on Azure, the Zuckerberg-led company intends to train the new model using its proprietary infrastructure, the sources said. 

The tech giant CEO also advocates for the new model to follow the same open-source approach as Meta’s previous AI offerings. This strategy would make the AI model freely accessible to companies, enabling them to construct AI-driven tools without incurring licensing fees.

Why It’s Important: Previously, it was reported that Apple Inc. (NASDAQ:AAPL) is also exploring AI offerings similar to ChatGPT and Bard, developing its own framework, known as ‘Ajax,’ and testing a chatbot referred to by some engineers as ‘Apple GPT.’

Image Credit – Pixabay

Check out more of Benzinga’s Consumer Tech coverage by following this link.

Read Next: Mark Zuckerberg’s Meta Must Defend Against Allegations Of Medical Privacy Breaches