Muse-Glimmer-30B-GGUF

Overview

Meta's Muse Glimmer 30B at the K-Quant-17GB cut - a dense multimodal agent model with a perception encoder, tuned for local tool use. Fits a 24 GB card with room for KV cache and the drafter.

Good for

• agentic
• chat
• coding

Recommended hardware

≈ 24 GB VRAM to run this 30B model at the K-Quant-17GB quantization.

Fits 24 GB+ consumer GPUs (e.g. RTX 3090, RTX 4090)

Links

• View on HuggingFace
• Run on Infersec