Meta eyes fruitful AI push in 2026 with new image, video, and text models

By TechWatch PH Staff

Meta Platforms is developing two new artificial intelligence models focused on visual and text-based generation, with a planned release targeted for the first half of 2026, according to reports.

The first model, code-named “Mango,” focuses on image and video generation, while the second, known internally as “Avocado,” is a text-based large language model.

Details about the models come from a report that cites an internal Q&A session led by Meta’s Chief AI Officer, where the company outlined its roadmap for upcoming AI projects.

Meta has publicly acknowledged its push into advanced AI, with the development of these models representing the next step beyond its existing offerings.

As of now, official technical specifications, capabilities, or confirmed product releases from Meta remain limited.

Future announcements from the company are expected to clarify when and how Mango and Avocado will be made available to developers and users.