It seems that Meta's new AI model development is more difficult than initially expected. According to multiple sources familiar with the matter, based on reports from The New York Times, the company's next-generation "Frontier" AI, currently under development, has not achieved satisfactory results in tests involving reasoning, programming, and text generation. Compared to models from competitors like Google, OpenAI, and Anthropic, there is a clear performance gap.

Meta CEO Mark Zuckerberg indicated in July of last year that the new AI model would be released within about a year. However, internally, it appears increasingly difficult to meet that deadline. Of particular interest is the model codenamed "Avocado." While this model demonstrated better performance than Meta's traditional AI systems, it still lagged behind Google's "Gemini 3.0," released in November of last year.

In response to this situation, Meta has decided to postpone the release of "Avocado," initially scheduled for April, to after May. The delay in deploying the AI model could be a significant setback in the highly competitive generative AI market.