Meta delays launch of flagship AI model Avocado until at least May 2026

Meta postponed the release of its new flagship AI model Avocado: instead of March 2026, the release is now expected no earlier than May. According to NYT and Reuters, the company decided to pause after internal evaluations where the model fell short of the best competitor systems.

Why the release was delayed

According to reports from Western publications, the problem is not that Avocado turned out to be weak in itself. In internal tests, it outperformed Meta's previous models, but did not show the leap that was expected within the company. Reuters sources say that in terms of answer quality, the model is currently between Google Gemini 2.5 and Gemini 3.0. For a project that was supposed to become the showcase of a new stage in Meta's AI strategy, this level turned out to be insufficient. Therefore, the company decided not to rush the release.

Instead of the March launch, the company has reportedly shifted the release window to no earlier than May, and according to some reports — even to June. Such a postponement well illustrates how fierce the race for frontier models has become: nowadays it's not enough to simply release "another strong model." You either need to reach the level of leaders or have a very clear advantage in price, speed, or product integration.

What is known now

From the published details, a picture emerges not of a failure, but of prolonged refinement. Meta is not scaling back the project and, on the contrary, continues to invest huge resources in AI — from computational infrastructure to its own chips. In January, the company said its annual capital expenditures could be between $115 billion and $135 billion. Against this backdrop, the Avocado postponement looks particularly sensitive: the larger the budget and expectations, the more noticeable each schedule shift becomes.

• The model was originally planned for release in March 2026.
• The new target is May 2026 or later.
• According to media reports, Avocado is already better than Meta's previous models, but falls short of fresh Gemini versions.
• Leadership discussed even temporary licensing of Gemini for Meta's AI products.

Particularly striking is the story about possible temporary Gemini licensing. There is reportedly no final decision on this option yet, but the discussion itself is telling. Meta has bet on its own AI ecosystem for many years, and here it admits a scenario where a competing model temporarily closes the product gap. This speaks to a more pragmatic approach: if the internal flagship is not ready, the business cares more about speed of rolling out features to users than ideology.

Pressure on Meta

Avocado is important not just as another large language model. For Meta, this is a test of a new stage of strategy in which the company is trying to close the gap with Google, OpenAI, and Anthropic not through individual demos, but through stable product releases. If the flagship has to be postponed because it doesn't look convincing enough against competitors, it strikes at two things at once: the team's internal pace and external perception of Meta's entire AI lineup.

"The next model will be good, but what matters more is that it shows our development trajectory."

This remark from a company representative well reflects the current mode: Meta is trying to win not one release, but a long series of updates throughout the year. But the market evaluates the nearest launch. If Avocado comes out later and without clear superiority, competitors get extra time to strengthen their models in search, assistants, coding, and enterprise scenarios. For Meta, this means that even a temporary pause turns into not just a delay, but a strategic risk.

What this means

The Avocado postponement shows a simple thing: even a company with a multi-billion AI budget can no longer afford to release a flagship "on a whim." The bar has already been set by Google, OpenAI, and Anthropic, and every new release is now compared not to the previous generation, but to the best models on the market right now.