Parameters
14B
Context Length
16K
Modality
Text
Architecture
Dense
KV Cache Rate
160 KB/tok
License
MIT License
Release Date
13 Dec 2024
Knowledge Cutoff
Nov 2024
API Pricing (per 1M)
Input: $0.07 · Output: $0.14
Precision
Context
Workload Memory: 10.27 GB
INT4 · 8K context
Base Weights: 6.52 GB
KV Cache: 1.25 GB
Runtime & Graphs: 2.50 GB
1x NVIDIA L4 24GB
10.3 / 21.6 GB usable
1x RTX 3060 12GB
10.3 / 10.8 GB usable
Apple M4 24GB
10.3 / 18.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Checkpoint format
FP16
CUDA
>= 11.8
PyTorch
>= 2.1
Transformers
>= 4.47.0
| Architecture | Precision | Attention kernel |
|---|---|---|
Blackwell, Hopper | Native FP16 | FlashAttention-3 |
Ada Lovelace, Ampere | Native FP16 | FlashAttention-2 |
Rank
#171
| Benchmark | Score | Rank |
|---|---|---|
Professional Knowledge | 0.704 | 42 |
Graduate-Level QA | 0.561 | 99 |
General Text | 1256 | 179 |
Intelligence Index | 0.06 | 310 |
General Knowledge Reference | 0.848 | 13 |
Overall Rank
#171
Coding Rank
-
Phi-4 is Microsoft's 14B parameter dense transformer model engineered with synthetic reasoning data for complex mathematics, logic, and scientific problem-solving. It delivers high-accuracy reasoning across a 16K token context window.
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
24
Key-Value Heads
8
Position Embedding
ROPE
RoPE Theta
250,000
Normalization
RMS Normalization
Activation Function
Swish
Dimensions
Base Precision
FP16
Hidden Dimension Size
3,072
Number of Layers
40
FFN Intermediate Size (Dense)
17,920
Tokenizer
Vocabulary Size
100,352
The Microsoft Phi-4 model family comprises small language models prioritizing efficient, high-capability reasoning. Its development emphasizes robust data quality and sophisticated synthetic data integration. This approach enables enhanced performance and on-device deployment capabilities.