Meta is eyeing to compete with Google and OpenAI’s growing dominance in the AI race by developing the series of next-generation artificial intelligence models.
According to The Wall Street Journal, the parent company of Facebook is in the process of generating an image and video-focused AI model named “Mango”.
Moreover, the social media giant is also working on a next-generation text-based large language model (LLM), named “Avocado” which would be equipped with advanced code capabilities.
Meta is expected to launch these AI-based models in the first half of 2026.
The development of these models will be the first major output from Meta Superintelligence Labs (MSL).
As per sources, Meta’s Mango model will be equipped with enhanced understanding of how the physical world operates via visual data analysis. In addition, the text to video feature of the tool will outshine all AI models.
If launched with advanced capabilities, Mango could be Gemini Nano Banana next big rival.
Recently, Gemini Nano Banana, Google’s latest image editing and generation tool, has taken the internet by storm due to its incredible features.
According to Meta’s Chief AI Officer Alexandr Wang, the company is also in the early stages of developing “world models”, aiming to understand physical reality by processing visual information.
