Meta has delayed the release of its next AI model, code-named Avocado, after internal results reportedly came in below the level needed to compete cleanly with the latest systems from Google and others.
Meta has pushed the release of its new AI model, Avocado, to at least May after initially targeting a March launch. Reuters says the model’s current performance sits between Google’s Gemini 2.5 and Gemini 3, which was enough to force a delay even as Meta keeps talking aggressively about its AI roadmap.
The timing is awkward. In January, Meta laid out capital spending of roughly $115 billion to $135 billion for the year in pursuit of what Mark Zuckerberg called “superintelligence,” and the company is still investing heavily in custom chips and model development. That makes the delay look less like a scheduling slip and more like a reminder that throwing money at foundation models does not guarantee frontier-level output on schedule.
Reuters also reports that Meta leaders have discussed the possibility of temporarily licensing Google’s Gemini to support some of its AI products, although no final decision has been made. If that conversation goes anywhere, it would be one of the clearest signs yet that even the largest AI players may need tactical stopgaps when internal model timelines slip.
