Why Meta Pushed Back the Avocado A.I. Model

Meta has delayed the planned public rollout of its next-generation A.I. model, codenamed Avocado, after internal performance testing found gaps in key tasks. An internal memo dated January 20 described Avocado as Meta’s most capable pre-trained base model to date, but engineers still flagged areas that need improvement. The pretraining phase is complete, yet follow-up fine-tuning and safety checks have been required before any broad release. The company has increased capital spending for A.I. infrastructure and will use additional compute and engineering time to improve results.

Meta’s approach appears to be shifting from open-source distribution toward a more controlled, closed model strategy, which could limit immediate access for developers. The launch window has been moved into the first quarter of 2026, with limited beta tests likely before a wider commercial offering. However, engineers will continue detailed evaluations of language, coding, and multimodal tasks before a broad rollout. Investors and product teams will watch how further testing changes the model’s capabilities and the company’s deployment plans.