Meta is developing a new AI model called Avocado that could be proprietary, marking a departure from its previous open-source approach with Llama. The model is being built inside Meta’s AI Superintelligence Labs under Chief AI Officer Alexandr Wang and is planned for release in 2026. Reports say internal issues with Llama 4, including delays and debate over whether to continue the project, contributed to the shift. Meta has also restructured parts of its AI organization, laying off workers in FAIR and seeing the departure of Chief AI Scientist Yann LeCun as Mark Zuckerberg increases spending to compete with OpenAI and Google.
