AI BriefingMeta Releases Muse Glimmer
Key point
Meta has released Muse Glimmer, a 30B multimodal model for local agents, under the Apache 2.0 license.
Details
Meta has released Muse Glimmer, a multimodal model with 30B parameters. It targets local agent use cases, including coding, document analysis, personal assistants, and privacy-centric applications.
The model is distributed under the Apache 2.0 license, enabling local execution and commercial use. Hugging Face added support for the following ecosystems simultaneously with the release:
transformersllama.cppvLLM- Inference Endpoints
In the published comparative benchmarks, Muse Glimmer achieved competitive results against Gemma4-31B and Qwen3.6-27B across several metrics, including MCP Atlas, DeepSearch QA, GAIA2, SWE-Bench Pro, SciCode, and Charxiv Reasoning. However, the Qwen series scored higher on certain benchmarks such as GDPval-AA, SkillsBench, OSWorld, and SWE-Bench Verified.
The model is available on the Hugging Face Hub at meta-models/Muse-Glimmer-30B.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Ollaya Launches Local Runtime for Open-Source Decision Models with TypeSafe SDK Compatibility