Meta has delayed the launch of its new artificial intelligence model, codenamed “Avocado,” until at least May. The company originally planned to release the AI model this month, but its performance currently sits between Google’s Gemini 2.5 and Gemini 3 models, prompting the postponement to May or June.
This delay comes despite Meta’s substantial investments in expanding its AI capabilities, including developing its own AI chips. In January, Meta announced capital expenditure plans of $115 billion to $135 billion for the year, aiming to achieve “superintelligence,” a state where AI surpasses human intelligence.
Meta has been working on “Avocado” for several months, but the model’s performance hasn’t quite matched the latest offerings from competitors. A Meta spokesperson addressed the situation, stating, “As we’ve said publicly, our next model will be good, but more importantly, show the rapid trajectory we’re on, and then we’ll steadily push the frontier over the course of the year as we continue to release new models.”
The spokesperson added, “We’re excited for people to see what we’ve been cooking very soon,” echoing CEO Mark Zuckerberg’s comments during an earnings call in January.
Interestingly, leaders within Meta’s AI division have even discussed the possibility of temporarily licensing Google’s Gemini to power their AI products. However, they have not made any final decisions on this front.
Back in December, media outlets reported that Meta was developing a text-based AI model called “Avocado” and expected to launch it in the first quarter of the year. This latest delay highlights the competitive and challenging landscape of the rapidly evolving AI industry, as companies strive to deliver cutting-edge technology. Meta’s continued investment signals its commitment to remaining a major player in the AI race, even if it means adjusting timelines to ensure quality and competitiveness.
