Muse Glimmer 30B — Hardware Requirements & GPU Compatibility
Vision
Muse Glimmer 30B is Meta's 30-billion-parameter dense vision-language model, purpose-built for local, always-on AI agents rather than general chat. It pairs a text decoder with a dedicated image encoder for reasoning over screenshots, charts, and documents, and includes native tool-calling with a separate reasoning channel so it can plan multi-step actions and recover from failures. At this size, local inference calls for quantization and a capable GPU; it fits a single high-end 24-32GB-class card, matching Meta's goal of running it entirely on consumer machines. The model supports a 131K token context window, enough for extended agent sessions. It is released under the Apache 2.0 license, allowing unrestricted commercial and research use. Published in August 2026, it is Meta's first open-weight model since Llama 4, and ships without the Llama licenses' monthly-active-user cap.
312.4K downloads
1.9K likesSpecifications
- Publisher
- Meta
- Family
- Muse Glimmer
- Parameters
- 29.8B
- Architecture
- MuseGlimmerForConditionalGeneration
- Context Length
- 131,072 tokens
- Vocabulary Size
- 202,048
- Release Date
- 2026-08-09
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Muse Glimmer 30B Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 13.1 GB | 24.3 GB | 12.66 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 13.5 GB | 24.7 GB | 13.03 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 15.0 GB | 26.2 GB | 14.52 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 15.4 GB | 26.5 GB | 14.89 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 18.3 GB | 29.5 GB | 17.87 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 21.7 GB | 32.9 GB | 21.22 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 25.0 GB | 36.2 GB | 24.57 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 30.3 GB | 41.4 GB | 29.78 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Muse Glimmer 30B?
Q4_K_M · 18.3 GBMuse Glimmer 30B (Q4_K_M) requires 18.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 24+ GB is recommended. Using the full 131K context window can add up to 11.2 GB, bringing total usage to 29.5 GB. 8 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroom
Which Devices Can Run Muse Glimmer 30B?
Q4_K_M · 18.3 GB41 devices with unified memory can run Muse Glimmer 30B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 Pro (24 GB).
Runs great
— Plenty of headroomDecent
— Enough memory, may be tightWhere to Download Muse Glimmer 30B
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Benchmarks
Benchmark details →Frequently Asked Questions
- How much VRAM does Muse Glimmer 30B need?
Muse Glimmer 30B requires 18.3 GB of VRAM at Q4_K_M, or 60.0 GB at BF16. Full 131K context adds up to 11.2 GB (29.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 29.8B × 4.8 bits ÷ 8 = 17.9 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 11.6 GB (at full 131K context)
VRAM usage by quantization
Q4_K_M18.3 GBQ4_K_M + full context29.5 GB- Can NVIDIA GeForce RTX 4090 run Muse Glimmer 30B?
Yes, at Q5_K_L (22.1 GB) or lower. Higher quantizations like Q6_K (25.0 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Muse Glimmer 30B?
For Muse Glimmer 30B, Q4_K_M (18.3 GB) offers the best balance of quality and VRAM usage. Q4_K_L (18.7 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 8.7 GB.
VRAM requirement by quantization
IQ2_XXS8.7 GBQ2_K13.1 GBIQ4_XS16.5 GBQ4_K_M ★18.3 GBQ5_019.1 GBBF1660.0 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Muse Glimmer 30B on a Mac?
Muse Glimmer 30B requires at least 8.7 GB at IQ2_XXS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Muse Glimmer 30B locally?
Yes — Muse Glimmer 30B can run locally on consumer hardware. At Q4_K_M quantization it needs 18.3 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Muse Glimmer 30B?
At Q4_K_M, Muse Glimmer 30B can reach ~262 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~36 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 18.3 × 0.65 = ~284 tok/s
Estimated speed at Q4_K_M (18.3 GB)
~284 tok/s~36 tok/s~284 tok/s~262 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Muse Glimmer 30B?
At Q4_K_M, the download is about 17.87 GB. The full-precision BF16 version is 59.55 GB. The smallest option (IQ2_XXS) is 8.19 GB.
- Which GPUs can run Muse Glimmer 30B?
8 consumer GPUs can run Muse Glimmer 30B at Q4_K_M (18.3 GB). Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX. 1 GPU have plenty of headroom for comfortable inference.
- Which devices can run Muse Glimmer 30B?
41 devices with unified memory can run Muse Glimmer 30B at Q4_K_M (18.3 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.