If Meta’s roadmap stays on track, both models are expected to launch in the first half of 2026. This move is more than just a product update, it signals Meta’s serious intent to reclaim influence in an AI landscape that has recently been shaped largely by ChatGPT, Gemini, and cutting edge video tools.
According to people familiar with the plans, Mango is being designed as a powerful visual model, focused on creating high quality images and videos.
Its goal is clear, to stand shoulder to shoulder with OpenAI’s Sora and Google’s latest visual AI systems. Meta wants Mango to be fast, creative, and reliable an everyday tool for modern digital storytelling.
Avocado, on the other hand, is a next generation language model. It is expected to handle writing, coding, and complex reasoning with greater depth and accuracy than Meta’s previous systems.
Insiders describe it as Meta’s most advanced language model so far, built around the idea of ‘world models’ AI systems that do not just predict words, but try to understand how the world works.
Leading this ambitious effort is Alexander Wang, Meta’s Chief AI Officer and the founder of Scale AI. After making a major investment in his company, Meta placed Wang in charge of its newly formed Meta Superintelligence Labs.
The mission of this lab is straightforward but bold, speed up research, sharpen innovation, and bring Meta back to the front line of AI development.
In 2025, Meta also reshaped its AI teams, bringing in dozens of top scientists and engineers from around the world, including former experts from OpenAI. This fresh talent, combined with heavy investment, shows that Meta is playing the long game.
All of this is happening while Google’s Gemini models and OpenAI’s tools continue to gain users at a rapid pace, raising the pressure across the industry. Meta, however, sees this pressure as fuel rather than a setback.
With Mango and Avocado, the company is sending a clear message: it is not watching the AI race from the sidelines. It is preparing a full-strength return.
If these plans succeed, 2026 could become a defining year for artificial intelligence, where the real competition is not just about faster models, but about vision, understanding, and creative intelligence.

