Meta Platforms is developing a new suite of artificial intelligence models, including an image and video generator codenamed “Mango” and a text-based model known as “Avocado,” The Wall Street Journal reported on Friday.
The initiatives were unveiled during an internal Q&A session by Chief Product Officer Chris Cox and Alexandr Wang, the Scale AI co-founder who recently took charge of the company’s new “superintelligence” lab.
The company targets a release window in the first half of 2026 for these next-generation tools.
The text-based “Avocado” model is being designed with a specific focus on improved coding capabilities, while the broader roadmap includes the exploration of “world models” capable of reasoning and planning without exhaustive training on every scenario.
This development push comes as Meta attempts to close the gap with competitors like OpenAI, Anthropic, and Google, following a year of significant restructuring that included the departure of Chief AI Scientist Yann LeCun.
Currently, Meta relies on the massive user base of its social platforms to drive adoption of its AI assistant, rather than a standalone hit product.
The upcoming models from Wang’s Meta Superintelligence Labs (MSL) represent a critical pivot to establish technical parity with industry leaders.
Key Takeaways:
- Meta is developing “Mango” (video/image) and “Avocado” (text) models for a planned release in the first half of 2026.
- Alexandr Wang leads the project under the new Meta Superintelligence Labs division.
- The roadmap focuses on improving coding proficiency and developing world models to compete with OpenAI and Google.
You may also want to check out some of our other recent updates.
Wanna know what’s trending online every day? Subscribe to Vavoza Insider to access the latest business and marketing insights, news, and trends daily with unmatched speed and conciseness! 🗞️
