Meta's new generation AI model Avocado has completed pre-training, internally referred to as the "most powerful" foundational model

Meta informs some employees that the internal code-named new model, Avocado, has become the company’s “most powerful pre-trained foundational model to date.”

On February 4th, according to The Information, Megan Fu, product manager at Meta’s Superintelligence Lab, disclosed new developments about the Avocado model in an internal memo dated January 20th.

The memo indicates that Avocado has completed the pre-training phase, which is the initial stage of AI model development where the model learns general knowledge, patterns, and relationships by exposure to large datasets.

The memo states that Avocado’s performance in testing surpasses leading open-source pre-trained foundational models. Although it has not yet undergone subsequent fine-tuning, the model is already competitive with top-tier post-training models in knowledge, visual perception, and multilingual capabilities.

This progress is significant for Meta’s competitive position in AI, but the actual performance of the model remains to be independently verified after public release. On Wednesday, Meta’s stock closed down 3.28%.

Superintelligence Lab Leading Development

The Avocado model was developed by Meta’s Superintelligence Lab, a team within Meta dedicated to AI research and development. This lab represents Meta’s latest organizational restructuring in the field of artificial intelligence.

Pre-training is a critical initial stage in AI model development, during which the model acquires general knowledge and patterns from large-scale datasets before being fine-tuned for specific tasks. Avocado has now completed this stage.

According to the internal memo, Avocado demonstrates competitiveness across multiple dimensions, achieved without post-training optimization. Post-training is the next phase in model development, typically used for tuning and improving for specific applications.

Although Meta is optimistic about Avocado’s capabilities internally, its actual performance cannot currently be independently verified. Only after the model is publicly released can external researchers and developers objectively assess its capabilities.

Reports indicate that Meta has not announced a release schedule or commercialization plan for the Avocado model. The company has also not disclosed whether the model will be open-sourced or used solely internally.

Risk Warning and Disclaimer

        The market carries risks; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users' specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Invest at your own risk.