Is Llama dead? Meta (META.US) will launch Avocado, a new AI model early next year

The Zhitong Finance App learned that according to people familiar with the matter, Meta (META.US) may release a new large-scale language model (LLM) in the first quarter of 2026 to further compete with companies such as Google (GOOGL.US) and OpenAI.

According to people familiar with the matter, the new model, codenamed “Avocado,” is regarded as Llama's successor — Meta has always faced challenges in developing the Llama model.

People familiar with the matter said Avocado will most likely be a proprietary model when it is released. Meta's current Llama series models are all open source models, allowing anyone to download and improve.

In June of this year, there was market news that Meta executives, including Zuckerberg, had discussed “cutting investment” in their Llama series language models and instead using models developed by competitors, including OpenAI and Anthropic products.

A few months later, Meta restructured its artificial intelligence division, partly to optimize its organizational structure to develop AI products faster to meet competition.

This summer, Meta spent nearly $15 billion to acquire shares in Scale AI and appointed its CEO Alexandr Wong as Meta's chief artificial intelligence officer.