Models

Meta launches Muse Glimmer for local agentic AI workflows

Meta has released Muse Glimmer, a 30-billion-parameter open-weight model designed for always-on local agents, offering reasoning, coding, multimodal understanding, tool use, and failure recovery on consumer hardware.

Meta has launched Muse Glimmer, a 30-billion-parameter open-weight model from Meta Superintelligence Labs designed for always-on agentic AI running locally.

Released under the Apache 2.0 license, the model supports multi-step reasoning, coding, reliable function calling, multimodal understanding, end-to-end task completion, and recovery from failed tool calls.

Meta optimized Glimmer to run on consumer hardware, including Macs and PCs equipped with a single GPU, reducing dependence on cloud infrastructure and internet connectivity. Developers can download the weights from Hugging Face and deploy the model through frameworks and services including Ollama, LM Studio, llama.cpp, vLLM, SGLang, and OpenRouter.

#
Meta

Read Our Content

See All Blogs
LLM Models

LLM testing of Grok 4.6: A cost-curve event, not a capability

Sarankumar S

August 14, 2026
Read more
LLM Models

Meta Muse Glimmer: Local agentic AI for enterprises

Vishesh Jain

August 13, 2026
Read more