Daijiworld Media Network - New York

New York, Oct 13: Meta Platforms Inc. on Monday unveiled its long-awaited artificial intelligence model, Muse Wonder, marking a major milestone in the company's multibillion-dollar effort to rebuild its AI division and compete with leading developers in the industry.

The new model, internally code-named “Watermelon”, is Meta's most capable AI model since CEO Mark Zuckerberg overhauled the company's AI operations through an aggressive recruitment drive that brought new talent into leadership positions.

Meta claims Muse Wonder represents a significant improvement over its earlier models, offering what the company describes as “frontier capabilities”. Its leadership says the model can compete with OpenAI's ChatGPT and Anthropic's Claude, particularly in coding and handling financial and legal queries.

However, comparing AI models remains challenging because different systems perform better at different tasks. Benchmarking results can also be manipulated and may not accurately reflect real-world performance. If Meta's claims withstand independent testing and Muse Wonder gains traction among researchers and developers, the launch could mark a significant turnaround for the company, whose AI models struggled to impress Silicon Valley last year.

“These are true frontier models,” Meta AI chief Alexandr Wang said in an interview on Monday. “We really think they are competitive with the best models out there, and particularly state of the art on personal and professional agentic work.”

Meta shares rose as much as 4.7% following the announcement. The stock had gained nearly 9% so far this year through Friday's close.

Over the past 18 months, Zuckerberg has intensified Meta's efforts to advance in the AI race, committing hundreds of billions of dollars to infrastructure needed to develop and operate advanced models. The company also recruited Wang to lead its new AI division as part of a $14-billion investment in Scale AI.

For Wang, the release of Muse Wonder marks the culmination of these efforts.

“This is the big moment,” he said.

Wang said Meta had sought to distinguish itself from competitors by developing large models designed for a broad range of consumer applications. Last month, the company introduced Muse, a personal AI agent designed to help users with everyday tasks and errands. Muse Wonder is expected to power the subscription version of the application soon.

“I think the rest of the ecosystem has been really focused on coding agents,” Wang said. “We do genuinely think that we've taken a differentiated focus to make the personal-agent use case really sing.”

A key part of Meta's strategy involves developing what it calls active memory and active focus capabilities. These features are designed to help AI agents retain relevant information about users over time and continue working on tasks as they become more complicated.

Wang said Muse Wonder could retain information across “days, weeks, months”, reducing the need for users to repeatedly explain their preferences and requirements. The agent would also be capable of trying alternative approaches when it encounters obstacles or reaches a dead end.

According to Wang, these capabilities would enable Meta's consumer AI agent to handle complex and even ambiguous tasks beyond straightforward requests.

Meta Superintelligence Labs (MSL), the division led by Wang, is already working on more advanced systems. He said the team had begun developing “larger and more scaled models” while simultaneously trying to reduce the cost of making the technology available to billions of users.

Muse Wonder is larger and more expensive to operate than Meta's smaller Muse Spark models, Wang said. The company is therefore developing specialised versions intended to retain much of Wonder's capabilities while operating efficiently enough for deployment on a massive scale.

For now, Muse Wonder is available to developers through the Meta Model API and Muse Code.

Asked how he would persuade sceptics who remain unconvinced that Meta has reached the forefront of AI development, Wang encouraged them to try the model and judge its capabilities for themselves.

He also dismissed criticism from detractors, saying he was not concerned about what he described as the “haters”.