Meta Open-Sources Muse Glimmer, a 30B Local Voice Agent Model

Meta released Muse Glimmer, an open-weight 30B-parameter agentic model optimized for on-device voice agents, running at about 20GB with 4-bit quantization on a consumer GPU and using speculative decoding to roughly triple decoding speed.

2026-08-19 ~ 2026-08-19 · 2 related posts

• Meta Releases Muse Glimmer: 30B Model Optimized for Always-On Local Voice Agents — Once_ina_Lifetime · 2026-08-19
• Meta Open-Sources Muse Glimmer: A 30B Local Agent Model with 3.1x Speed — dl_weekly · 2026-08-19