December 22, 2025 – According to The Wall Street Journal, Meta is fully mobilizing its workforce to develop a new generation of artificial intelligence models at its super – intelligence lab, which is led by Alexandr Wang, the co – founder of Scale AI.
At present, the company is advancing two new projects. One is an image and video model codenamed “Mango”, and the other is a brand – new text model internally known as “Avocado”.
It was noted that the media cited the content from an internal Q&A session held by Meta on Thursday. During this session, the tech giant revealed its plan to launch these new models in the first half of 2026. Alexandr Wang, along with Chief Product Officer Chris Cox, jointly unveiled this new development roadmap. Wang stated that Meta aims to enhance the code – generation capabilities of the text model. At the same time, it is working on developing new world models that can understand visual information and possess reasoning, planning, and action capabilities. Moreover, these models do not require pre – training for all possible scenarios.
In the global AI race, Meta has recently fallen behind competitors such as OpenAI, Anthropic, and Google. The company’s AI department has undergone a large – scale restructuring this year, which includes leadership changes and the recruitment of scientific research talents from other top – tier enterprises. However, several researchers who had joined Meta’s Super Intelligence Lab (MSL) have since left.
Just last month, Meta’s Chief AI Scientist Yann LeCun also announced his departure, planning to start his own startup.
So far, Meta has not yet created an AI product with dominant market power. The impressive user data of its AI assistant is entirely reliant on the company’s existing social network ecosystem, which has billions of users. This is because Meta has directly embedded the assistant into the search bars of its various applications.
This implies that the first batch of projects and models incubated by Meta’s Super Intelligence Lab bear the heavy responsibility of determining the success or failure of the company’s AI strategy.
