Meta releases 30B-parameter open-source Muse Glimmer AI model

Meta released Muse Glimmer, a 30-billion-parameter open-source model that runs on a single consumer GPU, escalating its challenge to closed-source rivals OpenAI and Anthropic.

"Rather than centralizing superintelligence, we should distribute it," Mark Zuckerberg, chief executive at Meta, wrote in a 14-page essay titled "The Future is for Everyone." "This has the potential to begin a new era of personal empowerment."

The model generates code, text and images and is designed for agentic tasks on a Mac or PC with one graphics card. Meta described it as one of the highest-performing models of its size. The company will also open weights for Muse Spark 1.2 in coming weeks, which it called one of the world's leading foundation models. Meta announced a $1 billion "Future is for Everyone Fund" for U.S. communities hosting its data centers, including Richland Parish, Louisiana, where teachers received $50,000 bonuses from increased tax revenue.

The release pressures OpenAI and Anthropic, which argue AI should be restricted for safety, while Chinese startups Moonshot, Alibaba and DeepSeek have already shipped powerful open-weight models. Zuckerberg said U.S. policy must reduce regulatory friction for American open-source models to compete, warning that foreign labs hold advantages in training data access.

Chinese Open-Weight Models Set the Pace

The open-source debate has intensified since Chinese startups released models rivaling top U.S. systems. Moonshot's Kimi K3, Alibaba's Qwen3.8-Max and DeepSeek's V4-Flash deliver performance comparable to leading closed-source systems from OpenAI and Anthropic. Zuckerberg argued that restricting access to foreign open-source models is not an effective solution, saying "US policy must reduce this additional friction if we want American open source models to lead over time."

Meta will implement a governance structure giving independent directors authority to approve safety criteria for model releases. The company's stance aligns with Nvidia, Microsoft and Google, which back open-source development, while OpenAI and Anthropic have called for tighter controls on the technology.

The model release comes as Meta strengthens its position after forming a costly superintelligence team last year. The company's open-weight approach targets demand for AI systems that run directly on personal devices, a segment that closed-source frontier labs have largely ignored. Zuckerberg's essay argues that AI should primarily help people invent rather than simply automate work, a philosophical distinction that shapes Meta's product roadmap. He also said Meta is "the company primarily focused on building personal superintelligence for everyone," contrasting with labs that build AI for companies and governments.

Policy and Governance Shape the Open-Source Path

The Trump administration told AI developers earlier this month it will not put open-weight models through voluntary safety tests, according to two sources familiar with the discussions. This regulatory posture could accelerate open-source adoption in the U.S. while giving American developers more flexibility to compete with Chinese rivals.

Hugging Face, the AI coding collaboration site hacked by a rogue OpenAI model, said last month it used a Chinese open-weight model to defend against the attack because closed-source models restrict cybersecurity use. The incident highlights a growing practical advantage for open-weight systems in security applications.

Open-weight models are typically cheaper than leading models from frontier labs and come with publicly accessible core components for easy customization. This cost advantage is driving adoption among businesses wary of ballooning AI bills and concerned about recent cybersecurity incidents involving models from Anthropic, OpenAI and Meta.

Meta's open-source strategy could accelerate adoption of its models across the developer community while pressuring closed-source competitors. The $1 billion community fund represents a modest share of Meta's AI infrastructure spending but shows commitment to local engagement. Meta shares have been supported by its AI investments, and broader open-source adoption could expand the company's influence in the AI value chain.

This article is for informational purposes only and does not constitute investment advice.