Meta delays launch of ‘Avocado’ AI model to May or later

Meta has postponed the launch of its upcoming artificial intelligence (AI) model, code-named Avocado, pushing the release timeline to May or later, according to reports.

The delay marks a shift from the company’s earlier plan to introduce the model this month. 

Sources told Reuters that the system’s performance currently falls between Gemini 2.5 and Gemini 3, advanced AI models developed by Google.

Despite the postponement, Meta continues to invest heavily in expanding its artificial intelligence capabilities. 

Earlier this year, the company outlined capital spending plans ranging from $115 billion to $135 billion, aimed at accelerating its AI development and advancing toward what executives describe as “superintelligence,” a theoretical stage where AI systems outperform human intelligence.

According to a report by The New York Times, Meta’s latest AI model has not yet matched the performance of competing systems from rival companies, prompting the decision to delay the launch.

A Meta spokesperson reiterated comments previously made by CEO Mark Zuckerberg during the company’s January earnings call.

“As we’ve said publicly, our next model will be good, but more importantly, it will show the rapid trajectory we’re on,” the spokesperson told Reuters. “We’ll steadily push the frontier over the course of the year as we continue to release new models.”

The spokesperson added that the company is “excited for people to see what we’ve been cooking very soon.”

The report also noted that leaders within Meta’s AI division have discussed the possibility of temporarily licensing Google’s Gemini technology to support some of its AI products, although no final decision has been made.