As artificial intelligence evolves beyond text-based interactions, companies across the tech industry are betting heavily on visual and multimedia models. Meta is also looking to catch up — having been left behind in the AI race thus far — developing an image and video generation model internally known as Mango, reports The Wall Street Journal. The model is expected to debut in 2026 and could eventually power creative tools across Meta’s social platforms. The move represents one of the company’s most ambitious efforts so far to remain competitive in the generative AI race.
Internally, Mango is positioned as a next-generation multimodal system, designed to generate and understand both still images and full-motion video. Unlike earlier AI image tools, which focus on producing single frames, video generation requires the model to maintain consistency across time – preserving characters, environments, lighting, and motion from one frame to the next. This significantly raises the technical bar, demanding larger datasets, more advanced architectures, and far greater computing resources.
The project is progressing along with another major internal initiative, a new large language model known as Avocado. While Avocado is aimed at text, reasoning, and coding tasks, Mango is intended to complement it on the visual side. The Mark Zuckerberg-led company’s interest in advanced video AI is closely tied to its core products. Platforms like Instagram and Facebook are already dominated by short-form and long-form video, and generative tools could significantly lower the barrier to content creation. AI-assisted video editing, automated visual effects, ad creation, and creator tools are all potential use cases.
The timing of the move becomes even more crucial as the social media giant elevates AI to a central priority while scaling back other major initiatives, including Reality Labs. At the same time, the company is facing significant financial pressure from its massive AI investments. Last month, the company announced plans to spend up to $600 billion on AI infrastructure across the US alone.
The development of Mango also comes as competition in AI video accelerates. Rivals, including OpenAI and Google, have made rapid progress in demonstrating video-generation systems, raising expectations for realism, duration, and controllability. As these technologies mature, the ability to generate high-quality video is increasingly seen as a key differentiator in the AI landscape.
The Tech Portal is published by Blue Box Media Private Limited. Our investors have no influence over our reporting. Read our full Ownership and Funding Disclosure →
