- Kyle Chua
Meta Reportedly Developing AI Model To Rival OpenAI's GPT-4
Updated: Dec 18, 2023
Meta's family of artificial intelligence (AI) models could soon add a new member.
According to the Wall Street Journal, the Facebook and Instagram parent company is developing a new large language model (LLM) that it expects to be as powerful as OpenAI's most advanced model yet, GPT-4. The model is also expected to be several times more powerful than the commercial version of Llama 2, another LLM that the social media giant launched in July of this year.
Llama 2 is Meta's answer to OpenAI's ChatGPT and Google Bard. It was released on Microsoft's Azure cloud infrastructure as a free-to-use, open source model, meaning anyone could use it for commercial or research purposes. Use cases of the model include the creation of chatbots for consumers and enterprises alike, language and code generation as well as building AI-powered tools.
Meta has reportedly been buying up Nvidia H100 chips that it'll need to beef up its infrastructure and start training the in-development model, which it plans to do so early next year. The acquisition of more chips would allow the company to train the model without needing to rely on Microsoft's Azure platform. Meta aims to have the AI model ready some time next year as well.
As for the model's use cases, the report claims it can help other companies build services that produce sophisticated text, analysis and other output of the sort.
The development of a new Meta LLM underscores the company's ambitions to advance AI tech as demand for it continues to grow and evolve. Many enterprises have started seeking out and adopting AI solutions to refine processes since OpenAI's viral chatbot ChatGPT became popular last year.
Apart from OpenAI and Google, Apple is rumoured to also be working on its own LLM model that's believed to be more powerful than GPT-4. The model is said to be built on the iPhone maker's own "Ajax" framework and powers a chatbot that's referred to internally as "Apple GPT".
Meta is reportedly developing a new large language model (LLM) that it expects to be as powerful as OpenAI's most advanced model yet, GPT-4.
The Facebook parent company has been buying up Nvidia H100 chips that it'll need to beef up its infrastructure and start training the in-development model, which it plans to do so early next year.
The model can help other companies build services that produce sophisticated text, analysis and other output of the sort.