DeepSeek-R1 8B
Parameters
8B
Context Length
64K
Modality
Text
Architecture
Dense
License
MIT License
Release Date
27 Dec 2024
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
System Requirements
VRAM requirements for different quantization methods and context sizes
1,024 tokens
Consumer
1x RTX 4090
24GB VRAM
Datacenter
1x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
64,000 tokens
Consumer
5x RTX 4090
24GB VRAM
Datacenter
2x NVIDIA A100
80GB VRAM
Apple Silicon
1x Apple M3 Max
128GB VRAM
Architecture Diagram
Evaluation Benchmarks
Rank
#169
| Benchmark | Score | Rank |
|---|---|---|
Graduate-Level QA | 0.49 | 104 |
Intelligence Index | 1 | 215 |
Rankings
Overall Rank
#169
Coding Rank
-
About DeepSeek-R1 8B
DeepSeek-R1 8B is a high-efficiency reasoning model distilled from DeepSeek-R1 onto open-weights backbones. It is optimized for low-hallucination logical deduction, function calling, and automated code generation.
Technical Specifications
Attention
Attention Structure
Multi-Layer Attention
Attention Heads
64
Key-Value Heads
64
Attention Head Dimension
-
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
SwigLU
Dimensions
Hidden Dimension Size
4,096
Number of Layers
40
FFN Intermediate Size (Dense)
14,336
Multi-Token Prediction Heads
-
Tokenizer
Vocabulary Size
128,256
About DeepSeek-R1
DeepSeek-R1 is a model family developed for logical reasoning tasks. It incorporates a Mixture-of-Experts architecture for computational efficiency and scalability. The family utilizes Multi-Head Latent Attention and employs reinforcement learning in its training, with some variants integrating cold-start data.