Meta AI’s smaller Llama 4 model will be ready “sometime early next year,” CEO Mark Zuckerberg has confirmed.
The billionaire businessman has boasted that the new language model for generative AI will be a “big deal on several fronts.”
“I expect that the smaller Llama 4 models will be ready first, and we expect [them] sometime early next year,” he said at the firm’s third quarterly earnings call, “and I think that they’re going to be a big deal on several fronts — new modalities, capabilities, stronger reasoning and much faster.”
The social media and tech mogul also revealed that they are training it on a GPU cluster “bigger than anything” his rivals are working on, to his knowledge.
“We’re training the Llama 4 models on a cluster that is bigger than 100k H100s or bigger than anything that I’ve seen reported for what others are doing,” he said.
However, Elon Musk had already said his AI startup, xAI, has worked with Nvidia on a cluster of the same amount.
Clusters are used to train and run complex AI models, such as those used in machine learning, deep learning, natural language processing (NLP) and computer vision.
