Meta Platforms Inc., the parent company of Facebook and Instagram, is reportedly delaying the launch of its new artificial intelligence model, named Avocado, by at least two months. The decision follows internal tests that revealed the model’s performance lagging behind competitors such as OpenAI, Google, and Anthropic, according to a report by the New York Times.
Initially scheduled for release this month, Avocado is now expected to debut around May. The internal assessments indicated that the new model fell short in key performance areas, including reasoning, coding, and writing. This setback is significant, as Mark Zuckerberg, Meta’s CEO, had previously indicated that the company’s AI advancements would “push the frontier” in the coming year.
In light of Avocado’s underperformance, leaders within Meta’s AI division are exploring a temporary licensing agreement with Google for its Gemini model. The decision underscores pressure on Meta to catch up in the rapidly evolving AI landscape, where competitors have set a high bar for performance.
Despite the disappointing results, the Times noted that Avocado outperformed Meta’s previous AI model and performed better than Google’s Gemini 2.5 released in March, although it fell short compared to Gemini 3.0 launched last November. This incremental improvement may not be enough to satisfy market expectations, particularly given the substantial resources Meta has dedicated to AI development.
Meta has invested billions in AI, hiring top talent and allocating $600 billion to build data centers to support these technologies. The company projected spending up to $135 billion this year alone, nearly double the $72 billion spent in the previous year on AI development. Such expenditures reflect Zuckerberg’s push to position Meta as a leader in AI innovation.
Meta’s chief AI officer, Alexandr Wang, who took on the role after the company invested $14.3 billion in the startup Scale AI in June, has been overseeing the development of Avocado. Wang is also leading TBD Lab, an internal AI lab responsible for the model’s development. The lab recently completed the pre-training phase of Avocado and began post-training in January, initially targeting a March release.
So far, the only product released by Wang’s team is Vibes, an AI video application akin to OpenAI’s Sora. However, the team has experienced some turnover, with several researchers departing before the launch of Avocado, raising questions about stability within the division.
Meta executives have reportedly debated whether Avocado should be open source, a model that would allow other developers to build upon its code. Historically, Meta has championed open-source technologies, arguing that they foster innovation. In contrast, companies like OpenAI and Anthropic have expressed concerns about the safety risks associated with making their models widely accessible.
Tensions have emerged within Meta’s leadership team, particularly between Wang and two of the company’s other top executives, Chris Cox and Andrew Bosworth. Disagreements have centered on how new AI models could enhance Meta’s advertising business. To address these challenges, Meta has announced the formation of an AI team under Bosworth, which will collaborate with Wang’s team.
As Meta navigates these internal and external pressures, its ability to innovate and adapt will be closely watched. The company’s commitment to advancing its AI capabilities is integral not only to its product offerings but also to its competitive positioning in the broader technology landscape. As the AI arms race intensifies, Meta’s success with Avocado and future models could have significant implications for its market presence and relevance in a rapidly changing industry.
See also
Amazon Implements AI Coding Oversight After 6.3 Million Lost Orders from Outages
Chinese DeepSeek AI Predicts XRP and Pi Coin Prices Could Surge After U.S.-Iran Ceasefire
Germany”s National Team Prepares for World Cup Qualifiers with Disco Atmosphere
95% of AI Projects Fail in Companies According to MIT
AI in Food & Beverages Market to Surge from $11.08B to $263.80B by 2032
