Meta joins AI price war! Launches first paid large model, outperforms Google, priced at only one-quarter of Anthropic.

In the increasingly competitive AI tools market, Mark Zuckerberg is betting on price to win.

Meta Platforms has released its most advanced AI model, Muse Spark 1.1, and for the first time launched a paid version for developers.

This is also the first time Meta has charged enterprises for model usage, opening a new revenue stream for the company. In an interview before the release, Zuckerberg said the product will be one of the most price-competitive AI models on the market.

"Since this is not an open-source model, I think this is the first time we've seriously launched an API service," Zuckerberg said.

"Our pricing will be very aggressive, very attractive."

Meta's stock opened lower but rose on Thursday, up 0.19% as of press time.

Focusing on AI Agent capabilities, priced at 1/4 of competitors

Zuckerberg said that the biggest improvement in Muse Spark 1.1 is its Agent capabilities.

Agent has become one of the hottest topics in the AI industry this year, typically referring to AI systems that can perform multi-step tasks on behalf of users. Zuckerberg stated that Muse Spark 1.1 has reached "industry-leading or very close to leading" levels in Agent reasoning and tool calling.

He also said the model has significantly improved its coding capabilities, with Meta's internal employees already using it to develop new products and features across the company's various applications.

At the same time, Meta is launching a new Meta Model API platform to charge developers.

Zuckerberg revealed that the pricing for Meta's API is about one-quarter of the official prices of similar top-tier models from OpenAI and Anthropic. Developers can use the model for free within a certain quota, and only pay after exceeding the specified Token limit.

Zuckerberg said Meta's goal is to make its AI technology accessible to more people.

"Some other AI labs have very high pricing and very high profit margins. We believe it's entirely possible to deliver cutting-edge, high-level intelligence at a much lower cost."

Committed to hundreds of billions of dollars, Meta's model surpasses Google's model for the first time

Zuckerberg, 42, is heavily investing in AI to catch up with competitors like OpenAI and Alphabet and achieve what he calls superintelligence — AI that can outperform humans across various tasks.

Meta has already committed to investing hundreds of billions of dollars in infrastructure needed to achieve superintelligence, including data centers and high-end AI chips. This week, the company also announced a $10 billion investment to build a data center in Canada and released a new image generation model.

Nevertheless, Meta's previous generations of models have lagged behind Anthropic, OpenAI, and Google in overall test scores.

However, Zuckerberg said that Muse Spark 1.1 has significantly improved its competitiveness, surpassing Google's Gemini model in multiple test items including Agent capabilities, coding, and multimodality.

"This is a very important milestone. This is the first time a Meta model has comprehensively surpassed all of Google's models."

AI lab restructured, strategy shifts to closed-source charging model

Over the past year, Zuckerberg has made significant adjustments to Meta to advance its AI strategy.

After a model release in the spring of 2025 underperformed expectations, Zuckerberg personally intervened to rebuild the AI team, including hiring Scale AI founder Alexandr Wang to lead a new AI department, and conducted multiple rounds of layoffs and organizational restructuring.

At the same time, Meta's AI strategy has undergone a major shift.

Previously, Meta insisted on building free, open-source models; now, the company is prioritizing the development of chargeable closed-source models, such as Muse Spark 1.1, and this transition essentially means rebuilding the entire technical system.

Zuckerberg expressed satisfaction with this. "Overall, our development is better than expected."

He acknowledged that Meta still lags behind leading AI labs like Anthropic and OpenAI, but revealed that the company is working on a new model codenamed Watermelon, which could help Meta further push the boundaries of AI intelligence.

Regarding the release timeline, he said the current focus remains on improving model quality, so he will not disclose a specific timetable.

Zuckerberg: Must master the underlying model to build the best AI assistant

Researching cutting-edge AI models is extremely costly, but Zuckerberg believes this investment aligns with Meta's long-term mission — to create a personal AI assistant that everyone can use.

He said that if Meta relies on other companies' models, it cannot guarantee that these models are designed around the core scenario of personal assistance.

"If you really want to build the best user experience, you have to be able to shape the underlying technology," Zuckerberg said, adding that mastering the underlying model allows Meta to "deliver what we believe is the best product experience."

He also disagreed with the view that AI will eventually become completely commoditized, with no difference between models.

Zuckerberg cited the example of Anthropic's latest model Mythos, which is restricted due to U.S. national security concerns, indicating that leading AI companies have begun to control the capabilities of their most advanced models rather than opening them fully.

"Truly advanced capabilities are not widely disseminated," he said. "Anthropic kept its strongest model and only released a relatively simplified version. Whatever their reasons, this at least shows that I don't think the industry will necessarily move toward making all the most advanced AI capabilities available to everyone."

Responding to commercialization doubts: Meta is building a complete AI monetization model

Meta expects capital expenditures to reach new highs in 2026, while investing billions of dollars to recruit AI talent for its Meta Superintelligence Labs, and plans to continue investing hundreds of billions of dollars in AI infrastructure.

Such massive spending has also raised external questions: how exactly will Meta make profits?

After the earnings report in April this year, despite revenue and profits exceeding expectations, the company's stock price fell nearly 9% due to market concerns about Meta's lack of a clear AI business model.

In recent months, Meta has gradually launched a series of commercialization initiatives, including:

Launching a consumer AI chatbot subscription service;

Planning to sell AI Agents to enterprises;

Starting to charge for API access to its flagship AI models;

Planning a cloud computing business, potentially selling computing power and AI infrastructure in the future.

Zuckerberg said Meta hopes to enable more developers to use advanced AI technology through more reasonable API pricing. "Someone has to develop these models and ensure that the highest quality intelligence can benefit everyone," he said.

He finally emphasized that Muse Spark 1.1 will also power the free version of Meta AI chatbot, meaning ordinary users can still experience this latest model for free.

Risk Disclaimer

        Market risk, invest with caution. This article does not constitute personal investment advice, nor does it consider the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment at your own risk.