Meta’s ambitious plan to launch a new flagship AI model has hit another snag. According to a report in The New York Times, the release, previously expected this spring, is now pushed to May. Internal sources indicate the model, internally referred to as 'Avocado,' has underperformed in key evaluations for reasoning, coding, and creative writing when measured against offerings from Google, OpenAI, and Anthropic.

This delay comes despite a staggering financial commitment. Last year, Meta spent billions recruiting top engineering talent and completed a $14.3 billion deal to bring Scale AI's Alexandr Wang on board as chief AI officer. The company has signaled even heavier spending ahead, projecting expenses up to $135 billion this year to fuel its AI labs. Yet, the technological lead remains with others. Google's Gemini 3, OpenAI's GPT-5 updates, and Anthropic's capable agentic tools have each set a brisk pace that Meta has struggled to match.

In a statement, a Meta spokesperson said, "Our next model will demonstrate the rapid trajectory we're on. We'll steadily push the frontier throughout the year and are excited to share our progress soon."

The company's most notable AI announcement this year wasn't a model, but an acquisition: Moltbook, a social network for AI bots. Meanwhile, Meta faces separate controversies. Its Ray-Ban smart glasses are central to a lawsuit alleging unauthorized recording, and its core social apps are under scrutiny in a major trial examining potential harms to teenagers.

Source: CNET