Muse Glimmer
Muse Glimmer is an open-weights language model from Meta, released August 10, 2026. It has 30B parameters (dense). The context window is 128K tokens.
Today, we're introducing Muse Glimmer, the next model from Meta Superintelligence Labs, and open sourcing the model weights under a permissive Apache 2.0 license. Muse Glimmer is a 30-billion-parameter model optimized for always-on local agent workflows.
Check it against your own PC →
Can you run Muse Glimmer locally?
The smallest common GPU it runs on is 12 GB of VRAM (with 32 GB of RAM): Q3_K_M, split GPU / CPU, roughly 5 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 5.2 GB | 7.1 GB | ~4 |
| 8 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 7.2 GB | 5.1 GB | ~5 |
| 12 GB VRAM | Runs, slowly | Q3_K_M (okay) | 11.2 GB | 4.3 GB | ~5 |
| 16 GB VRAM | Runs on GPU | IQ3_XXS (noticeable loss) | 12.3 GB | – | ~26 |
| 24 GB VRAM | Runs on GPU | Q5_K_M (excellent) | 22.1 GB | – | ~26 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 25.3 GB | – | ~42 |
Ready-made quantizations on Hugging Face: unsloth/Muse-Glimmer-30B-GGUF · meta-models/Muse-Glimmer-30B-GGUF · bartowski/Muse-Glimmer-30B-GGUF · AtomicChat/Muse-Glimmer-30B-GGUF · 0bserverx/Muse-Glimmer-30B-Heretic-Uncensored-GGUF · ggml-org/Muse-Glimmer-30B-GGUF
Benchmark results
Epoch AI has not published results for Muse Glimmer yet. New models are usually evaluated within days to weeks.
Links
Announcement · Hugging Face: meta-models/Muse-Glimmer-30B
More from Meta
Muse Ember · Muse Spark Open · Muse Wonder · Muse Wonder Full · Muse Spark 1.3 · Muse Spark 1.3 Contributor · Muse Glimmer 30B assistant · Muse Spark 1.2