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

Meta's introduction of the Muse Wonder models represents a clear strategic shift towards developing frontier AI, aiming to compete directly with companies like OpenAI and Anthropic. This ambition is underscored by the company's US$14 billion investment in Scale AI in June 2025 and the subsequent appointment of Scale AI's co-founder, Alexandr Wang, as Meta's chief AI officer, heading its new Superintelligence Labs division.

The performance claims for the Wonder models are supported by benchmarking data from Artificial Analysis, which Meta cited as showing its new models surpass its own earlier Spark models and various offerings from competitors. However, Artificial Analysis's website also indicates that the Wonder models are among the most expensive available, a factor that could influence their practical application and broader market adoption.

This initiative aligns with Meta's broader financial strategy to diversify revenue beyond advertising, a sector from which it has historically derived most of its income. The company has committed substantial resources, with AI-related investments potentially reaching US$145 billion this year, reflecting a determination to establish a strong position in the AI market, partly through a paid subscription model for its Muse personal agent app.