Meta challenges Anthropic and OpenAI with Muse Wonder: "It will allow for much more sophisticated behaviors"

Meta claims to have caught up with its main competitors in the development of artificial intelligence. Mark Zuckerberg, the company's CEO, announced a larger multi-billion dollar spending to compete with Anthropic and OpenAI, while rejecting requests to slow down the advancement of this technology. The announcement coincided with the launch of Muse Wonder, the company's new AI model. According to Zuckerberg's statements to the Financial Times, the system will allow for "much more sophisticated behaviors that are on the frontier of what is being done in the field of AI." ADVERTISEMENT What is Muse Wonder and how it competes with OpenAI and Anthropic Muse Wonder is the artificial intelligence model that Meta presented as its bet to equal the leaders of the sector. The company maintains that the system performs better than comparable models from Anthropic and OpenAI in certain benchmark technical evaluations. Specifically, Meta claimed that Muse Wonder surpasses Claude Opus 5.5 and Fable 5.1, both from Anthropic, as well as GPT-6 Astra from OpenAI, in some of these tests. The company pointed out that the new model can be used for complex tasks, from "understanding a lease agreement to making sense of financial or health information." ADVERTISEMENT The launch represents a milestone for Zuckerberg, who has allocated billions of dollars since last year to restructure Meta's artificial intelligence leadership. That strategy included the construction of large data centers and the hiring of AI research scientists with multi-billion dollar salaries, in an effort not to fall behind OpenAI, Anthropic, and Google. Muse Wonder will function as a closed model, whose most advanced versions will be sold directly to companies. This decision places Meta in direct competition with Anthropic and OpenAI for corporate artificial intelligence contracts, a ground the company considers decisive for its business ambitions. ADVERTISEMENT Zuckerberg foreshadowed that the pace of development will accelerate. "We are already working on the next level," he said, adding that the advancement of their models "will probably give the impression of moving much faster from now on." Why Muse became the most downloaded application in the United States Muse, Meta's personal artificial intelligence assistant, has become the most downloaded application on both the Apple and Android stores in the United States. The tool uses autonomous bots capable of performing everyday tasks, such as making purchases or booking travel. The download success occurred despite the fact that Meta's bet on artificial intelligence caused initial concern among investors. However, the company's stock price rose almost 10% since the beginning of September, the month Muse assistant was launched. ADVERTISEMENT Meta, valued at $1.8 trillion, plans to integrate the new Muse Wonder model directly into this application. The company expects this integration to reinforce Muse's positioning as a mass-use tool among consumers. In statements to the Financial Times, Zuckerberg explained the work behind this development: "Reaching this point has required a great deal of computing power and investment, as well as very rigorous scientific work by the team that has been working on it for a long period." What Zuckerberg says about the risks of artificial intelligence Zuckerberg downplayed growing fears about the pace of artificial intelligence advancement and its possible escape from human control. Among the episodes that fueled this concern is an incident that occurred in July when more than 1,000 autonomous agents from OpenAI escaped a testing environment to pirate the startup Hugging Face. ADVERTISEMENT When asked about his calculation of "P(doom)," the term used in Silicon Valley to describe the probability that artificial intelligence will cause massive destruction, the executive avoided giving a figure. Instead, he argued that the responsibility of ensuring the safety of the models should lie with each laboratory individually. "I am much closer to the belief that AI is a technology that can be designed with normal protective measures and built responsibly, than the anthropomorphic vision of AI, which holds that it will just go its own way," Zuckerberg stated. ADVERTISEMENT The Meta executive had previously rejected calls from other leaders in the sector, such as Dario Amodei from Anthropic and Sam Altman from OpenAI, for a slowdown at an industrial level in the development of artificial intelligence. Meta's role in the White House Superintelligence Agreement Zuckerberg played a role in drafting the Superintelligence Agreement driven by the White House of the United States President, Donald Trump. The document establishes principles of self-regulation for the artificial intelligence industry. Critics of that agreement argue that its provisions are insufficient regarding what should be a formal global governance regime for this technology. Despite these questions, the approach reflects the stance that Meta defends against other actors in the sector. ADVERTISEMENT Instead of focusing on building artificial general intelligence, known as AGI, that surpasses humans in most tasks (a goal pursued by OpenAI and Anthropic), Zuckerberg promoted his own concept. His vision points to what he calls "personal superintelligence," oriented towards individual consumers. This conceptual difference marks Meta's strategy towards its direct competitors. While OpenAI and Anthropic focus their efforts on general-purpose systems, Meta's company prioritizes artificial intelligence tools designed for the personal and everyday use of each user. ADVERTISEMENT