Meta Platforms (NASDAQ: META) launched a new open-weight artificial intelligence model Monday, with CEO Mark Zuckerberg simultaneously calling for American policymakers to reduce barriers facing domestic AI developers.
The new model, named Muse Glimmer, is considerably smaller than flagship AI systems produced by leading competitors and has been designed primarily for agentic tasks.
Unlike large frontier models requiring major cloud infrastructure, Muse Glimmer can operate directly on a Mac or PC equipped with a single graphics card.
Meta indicated that additional open-weight models are expected to follow in the near future, signaling a broader push into accessible, device-ready AI technology.
Open-weight models generally offer a less expensive alternative to closed systems developed by frontier AI laboratories such as OpenAI and Anthropic, while giving developers greater flexibility to customize applications.
Support for open-weight technology has grown as businesses become increasingly sensitive to rising AI costs and cybersecurity concerns involving models from Anthropic, OpenAI and Meta.
Hugging Face, an AI development and collaboration platform that was hacked by a rogue OpenAI model, said last month that it used a Chinese open-weight model to defend against the incident because closed-source systems restrict their use for cybersecurity-related tasks.
Meta shares, which have declined around 10% since the beginning of the year, gained approximately 1% in pre-market trading on Monday following the announcement.
Chinese technology companies currently hold prominent positions in the open-weight segment, with models including Moonshot’s Kimi K3, Alibaba’s Qwen3.8-Max and DeepSeek’s V4-Flash delivering competitive performance against leading American AI systems.
Zuckerberg argued that U.S. policymakers must act if American developers are to lead in open-weight AI, pointing to regulatory disadvantages facing domestic labs relative to their Chinese counterparts.
“Foreign labs currently hold several advantages here since American labs have to comply with many additional restrictions on training data,” Zuckerberg said, referring to open-source models.
“U.S. policy must reduce this additional friction if we want American open source models to lead over time,” Zuckerberg added, while arguing that restricting access to foreign open-source systems would not provide an effective solution.
The Trump administration reportedly told AI developers earlier this month that open-weight models would not be subjected to voluntary government safety testing, according to people familiar with the discussions.
Zuckerberg also expressed support for AI model distillation, a technique in which a powerful AI system’s capabilities are used to train a smaller and potentially more efficient model, which could prove particularly relevant as Meta develops systems for consumer hardware.
Alongside its open-weight strategy, Meta plans to establish a governance framework giving its independent directors authority to approve safety criteria applied when determining whether models should be released.
The Muse Glimmer launch reflects Meta’s broader effort to combine smaller, more accessible AI models with a push for regulatory changes that Zuckerberg believes could strengthen U.S. developers’ ability to compete with increasingly capable Chinese open-weight systems.
