Meta unveiled a new open-weight model on Monday, with CEO Mark Zuckerberg advocating for fewer curbs for open-source AI models in the United States, in order to allow the country's tech players to compete with Chinese rivals.
A 6,500-word essay by CEO Mark Zuckerberg supporting the open-source ecosystem over closed-source competitors has been released alongside Meta's new open-weight AI model, Muse Glimmer. The new model was created especially for consumer gadgets and agentic duties.
In an attempt to meet the demand for AI systems that operate directly on people's devices, the new model, Muse Glimmer, is significantly smaller than top AI models from competitors and is instead made for agentic tasks. It can run on a Mac or PC with a single graphics card.
This comes days after the recent demonstration of support for open-weight AI by Big Tech firms. The concept is gaining popularity as companies become more cautious about skyrocketing AI costs and are concerned about recent cybersecurity breaches involving models like Anthropic, OpenAI, and Meta.
ALSO READ: China's Open-Weight Models Will Be Spared US Safety Tests
Notably, leading models from frontier laboratories like OpenAI and Anthropic are usually more expensive than open-weight models. Unlike closed models, which businesses maintain complete control over, open-weight models also include publicly accessible essential components that enable simple adaptation.
In a statement, Zuckerberg argued that if domestic companies were to take the lead in open-weight models, the United States needed to reconsider its policy.
The competition for open-weight models is being led by Chinese startups, such as Moonshot's Kimi K3, Alibaba's Qwen3.8-Max, and DeepSeek's V4-Flash, which provide performance comparable to the best systems produced by American AI laboratories. In contrast, the top models produced by American companies OpenAI, Anthropic, and Alphabet's Google are proprietary.
In reference to open-source models, Zuckerberg stated, "Foreign labs currently hold several advantages here since American labs have to comply with many additional restrictions on training data."
Zuckerberg stated that limiting access to foreign open-source models was ineffective and that "US policy must reduce this additional friction if we want American open source models to lead over time."
Zuckerberg also supported AI model distillation in his statement, which is the process of employing a strong AI system to train a smaller model. According to him, Meta would put in place a governance framework that would allow its independent directors to accept the safety standards for model releases.
ALSO READ: Can China Overtake US In AI? Here's What Elon Musk Thinks
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.