Llama 4 Scout
Active Parameters
109B
Context Length
10M
Modality
Multimodal
Architecture
Mixture of Experts (MoE)
License
Llama 4 Community License Agreement
Release Date
6 Apr 2025
Knowledge Cutoff
Aug 2024
System Requirements
VRAM requirements for different quantization methods and context sizes
1,024 tokens
Consumer
12x RTX 4090
24GB VRAM
Datacenter
4x NVIDIA A100
80GB VRAM
Apple Silicon
3x Apple M3 Max
128GB VRAM
10,000,000 tokens
Consumer
281x RTX 4090
24GB VRAM
Datacenter
65x NVIDIA A100
80GB VRAM
Apple Silicon
63x Apple M3 Max
128GB VRAM
Architecture Diagram
Evaluation Benchmarks
Rank
#149
| Benchmark | Score | Rank |
|---|---|---|
QA Assistant | 0.873 | 17 |
Summarization | 0.684 | 22 |
General Knowledge | 0.796 | 22 |
StackUnseen | 0.16 | 31 |
Professional Knowledge | 0.70 | 39 |
General Text | 1321 | 136 |
Rankings
Overall Rank
#149
Coding Rank
#105
About Llama 4 Scout
Llama 4 Scout is a multimodal Mixture-of-Experts model from Meta activating 17B parameters across a massive 10M token context window. Optimized to run on a single H100 with Int4 quantization, it excels at multi-document analysis and repository synthesis.
Technical Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
64
Key-Value Heads
8
Attention Head Dimension
128
Position Embedding
ROPE
RoPE Theta
500,000
Sliding Window Attention
No
Sliding Window Size
-
Sliding Window Ratio
-
Linear Attention
-
Linear Attention Ratio
-
Normalization
RMS Normalization
Activation Function
Swish
Dimensions
Hidden Dimension Size
8,192
Number of Layers
80
FFN Intermediate Size (Dense)
8,192
Multi-Token Prediction Heads
-
Tokenizer
Vocabulary Size
202,048
Mixture of Experts
Total Expert Parameters
17.0B
Number of Experts
16
Active Experts
2
Shared Experts
-
FFN Intermediate Size (per Expert)
-
Dense Layers Before MoE
-
About Llama 4
Meta's Llama 4 model family implements a Mixture-of-Experts (MoE) architecture for efficient scaling. It features native multimodality through early fusion of text, images, and video. This iteration also supports significantly extended context lengths, with models capable of processing up to 10 million tokens.