Meta says its new AI models on par with OpenAI, Anthropic

Meta, the parent company of Facebook and Instagram, announced on Monday the launch of two new frontier artificial intelligence models named Wonder and Wonder Full. These models, developed by the company's newly formed Superintelligence Labs division, are claimed by Meta to be on par with AI models from OpenAI and Anthropic.

Meta CEO Mark Zuckerberg appointed Alexandr Wang as the company's chief AI officer, overseeing the Superintelligence Labs division. Wang, who was previously a co-founder of Scale AI, a San Francisco-based AI startup, was brought on board after Meta made a $14 billion investment in the company in June 2025.

Wonder is Meta's first new model since April, when the company launched a frontier model called Spark. The new models can be used in Meta's new personal agent app, Muse, which is currently available to paid subscribers only. Wang posted on X, a social media platform owned by Meta, that Wonder is the company's largest and most capable model yet and rivals the big AI labs, OpenAI and Anthropic.

In a blog post, Meta shared benchmarking data from an independent firm called Artificial Analysis, which suggests that Wonder outperforms Meta's previous models, Spark, and models developed by Anthropic, OpenAI, and Google. However, Artificial Analysis' website notes that Wonder is also among the most expensive models.

Meta has been investing heavily in AI, with Zuckerberg stating that the company's spending on AI-related investments could reach $145 billion this year. Despite growing concerns about the safety and security of AI, the company continues to compete in the global race to dominate the technology. Meta currently has over 3.6 billion daily active users worldwide.