Mark Zuckerberg Announces Meta’s Muse Glimmer AI Model
Meta has released Muse Glimmer, a 30-billion-parameter open-weight AI model designed for local, offline use on consumer hardware.
Mark Zuckerberg Founder, Chairman and Chief Executive Officer - META
Opinions expressed by Entrepreneur contributors are their own.
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
Meta has released Muse Glimmer, a 30-billion-parameter open-weight AI model designed to run locally on consumer hardware, including a Mac or PC with a single graphics card. The model is built for agentic workflows, enabling it to handle tasks such as coding, tool use, and other multi-step computer-based operations without requiring continuous access to cloud infrastructure.
The model’s weights have been released under the Apache 2.0 licence, allowing developers to download, run, and modify it freely. According to Meta, Muse Glimmer supports text and image inputs, local coding, function calling, and LLM-as-a-judge evaluation, and is designed to operate entirely offline once deployed.
The model features a 131K-plus context window and supports more than 100 languages, with a focus on local agentic workflows involving multi-step reasoning, tool use and failure recovery. Muse Glimmer is designed to run on consumer hardware, eliminating the need for the large-scale computing infrastructure typically required by cloud-based AI models.
Meta said Muse Glimmer is built for always-on local agents and is available through Hugging Face. It supports local deployment tools and frameworks such as llama.cpp, Ollama, LM Studio, vLLM, SGLang, ExecuTorch and MLX.
The company has also announced that it will soon release the weights for Muse Spark 1.2, its latest and more capable AI model. The Muse Glimmer launch is accompanied by a 6,500-word essay by Meta CEO Mark Zuckerberg, titled “The Future is for Everyone.”
In the essay, Zuckerberg argues that advanced AI should not be controlled by a small number of companies or institutions. He said access to AI should be distributed more widely and argued that open-weight models give users and developers greater control over how AI systems are used.
“The notion that AI is so dangerous that the only safe path is an extreme concentration of power seems inherently problematic,” Zuckerberg adds.
The essay sets out Zuckerberg’s broader idea of ‘personal superintelligence’, where AI systems are available to individuals and can act on their behalf. He also discusses AI regulation, data centres, employment, surveillance, cybersecurity and national security.
Meta has also announced a USD 1 billion fund for communities affected by the expansion of its data-centre infrastructure. Zuckerberg said the fund would support communities where Meta develops data centres and AI infrastructure. The announcement marks Meta’s return to releasing open-weight models as it continues developing its Muse AI model family through Meta Superintelligence Labs.
Meta has released Muse Glimmer, a 30-billion-parameter open-weight AI model designed to run locally on consumer hardware, including a Mac or PC with a single graphics card. The model is built for agentic workflows, enabling it to handle tasks such as coding, tool use, and other multi-step computer-based operations without requiring continuous access to cloud infrastructure.
The model’s weights have been released under the Apache 2.0 licence, allowing developers to download, run, and modify it freely. According to Meta, Muse Glimmer supports text and image inputs, local coding, function calling, and LLM-as-a-judge evaluation, and is designed to operate entirely offline once deployed.
The model features a 131K-plus context window and supports more than 100 languages, with a focus on local agentic workflows involving multi-step reasoning, tool use and failure recovery. Muse Glimmer is designed to run on consumer hardware, eliminating the need for the large-scale computing infrastructure typically required by cloud-based AI models.
