Muse Glimmer 30B
Parameters
30B
Context Length
131K
Modality
Multimodal
Auxiliary Parameters
1.5B
Architecture
Dense
License
Apache 2.0
Release Date
9 Aug 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Input: $0.30 · Output: $1.20
System Requirements
VRAM requirements for different quantization methods and context sizes
1,024 tokens
Consumer
3x RTX 4090
24GB VRAM
Datacenter
1x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
131,072 tokens
Consumer
3x RTX 4090
24GB VRAM
Datacenter
1x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
Architecture Diagram
Evaluation Benchmarks
Rank
#104
| Benchmark | Score | Rank |
|---|---|---|
Graduate-Level QA | 0.835 | 51 |
Coding Index | high 0.49 | 90 |
Agentic Index | high 0.09 | 93 |
Web Development | 1355 | 96 |
General Text | 1425 | 96 |
Intelligence Index | high 0.06 | 305 |
Rankings
Overall Rank
#104
Coding Rank
#82
About Muse Glimmer 30B
Muse Glimmer 30B is a dense open-weight multimodal model distilled from Meta's Muse Spark architecture. It is optimized to execute long-horizon autonomous agent tasks efficiently on consumer-grade hardware.
Technical Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
32
Key-Value Heads
2
Attention Head Dimension
128
Position Embedding
ROPE
RoPE Theta
500,000
Sliding Window Attention
Yes
Sliding Window Size
2,048
Sliding Window Ratio
75.0%
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Auxiliary Parameters
1.5B
Hidden Dimension Size
6,656
Number of Layers
52
FFN Intermediate Size (Dense)
19,968
Tokenizer
Vocabulary Size
202,048
Resources
About Muse Spark
The Muse Spark model family developed by Meta.