Meta's 'Avocado' AI Model Faces Delays Amid Strategic Shifts

Meta Platforms is adjusting its artificial intelligence strategy, with the release of its forthcoming AI model, codenamed 'Avocado,' now anticipated in early 2026, a shift from the initial late 2025 timeline. This development follows the establishment of Meta Superintelligence Labs (MSL) in June 2025, a division dedicated to advancing AI capabilities.

The 'Avocado' model is expected to be a proprietary system, marking a departure from Meta's previous open-source approach with its Llama models. In July, CEO Mark Zuckerberg indicated a more cautious stance on open-sourcing advanced AI models, stating, 'We’ll need to be rigorous about mitigating these risks and careful about what we choose to open source.'

Meta's AI initiatives have also seen significant leadership changes. The company invested $14.3 billion in Scale AI and appointed its CEO, Alexandr Wang, to lead MSL. Additionally, Chief AI Scientist Yann LeCun announced his departure to start a new AI research company, focusing on developing advanced AI systems with capabilities such as understanding the physical world and reasoning.

A Meta spokesperson commented on the 'Avocado' model's progress, stating, 'Our model training efforts are going according to plan and have had no meaningful timing changes.'

These developments underscore Meta's strategic realignment in AI development, emphasizing proprietary models and leadership restructuring to enhance its position in the competitive AI landscape.