MENLO PARK, Calif. — Meta has launched Muse Glimmer, a new open-weight artificial intelligence model that signals the company's renewed push to make advanced AI technology more accessible to developers and consumers.

Unlike some of the industry's largest AI systems that depend heavily on cloud infrastructure, Muse Glimmer is designed to operate on consumer computers equipped with a single graphics card, potentially making sophisticated AI capabilities more accessible outside major technology companies and data centers.

The model is built for agentic tasks, allowing it to handle multi-step activities rather than simply generating conventional text responses. This could make the technology useful for coding, research, administrative work and other computer-based tasks.

Meta's move comes as competition among U.S. technology companies intensifies over the future of artificial intelligence. Companies including OpenAI, Google, Anthropic and Meta are investing heavily in increasingly capable AI systems while debating whether advanced models should remain closed or have their weights made available to developers.

Meta CEO Mark Zuckerberg has continued to advocate for wider access to AI technology. In connection with the launch, he argued that the United States needs policies that allow American developers to compete effectively with rapidly advancing AI companies in China.

The company is also expected to release the weights for a more powerful Muse Spark model, potentially giving developers access to another layer of Meta's AI technology.

Meta's broader Muse strategy already includes Muse Spark, which the company introduced as a multimodal reasoning model for its AI products. Meta says Muse Spark supports tool use, visual reasoning and multi-agent orchestration.

The latest release could have significant implications for the AI industry because smaller models that can operate locally may reduce dependence on expensive cloud computing. Developers could potentially customize AI systems for specialized applications while keeping more processing on their own hardware.

At the same time, open-weight AI models raise continuing questions about cybersecurity, misuse, safety testing and responsible deployment. The availability of model weights can give researchers greater freedom to inspect and adapt AI systems, but it can also make certain safeguards harder to control centrally.

Meta's decision therefore places the company directly in the growing debate over how much control technology companies should maintain over increasingly capable artificial intelligence.

Muse Glimmer's launch could become an important test of whether powerful AI can move from centralized cloud platforms toward more affordable, locally operated computing.