Qwen2.5-14B
Parameters
14B
Context Length
131K
Modality
Text
Architecture
Dense
KV Cache Rate
80 KB/tok
License
Apache 2.0
Release Date
19 Sept 2024
Knowledge Cutoff
Jun 2024
API Pricing (per 1M)
Self-hosted only
Hardware & System Requirements
Precision
Context
Workload Memory: 9.64 GB
INT4 · 8K context
Base Weights: 6.52 GB
KV Cache: 0.63 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
9.6 / 21.6 GB usable
1x RTX 3060 12GB
9.6 / 10.8 GB usable
Apple M4 24GB
9.6 / 18.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Runtime requirements
Checkpoint format
FP16
CUDA
>= 11.8
PyTorch
>= 2.1
Transformers
>= 4.43.1
| Architecture | Precision | Attention kernel |
|---|---|---|
Blackwell, Hopper | Native FP16 | FlashAttention-3 |
Ada Lovelace, Ampere | Native FP16 | FlashAttention-2 |
Architecture Diagram
Rankings
Overall Rank
#195
Coding Rank
-
About Qwen2.5-14B
Qwen2.5-14B is a mid-tier dense foundation model from Alibaba Cloud delivering frontier-grade reasoning, coding, and mathematical problem-solving. Pretrained on 18T tokens, it supports instruction-following and tool use across a 131K context window.
Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
80
Key-Value Heads
8
Position Embedding
ROPE
RoPE Theta
1,000,000
Sliding Window Attention
No
Sliding Window Size
131,072
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Base Precision
FP16
Hidden Dimension Size
5,120
Number of Layers
40
FFN Intermediate Size (Dense)
13,824
Tokenizer
Vocabulary Size
152,064
About Qwen2.5
Qwen2.5 by Alibaba is a family of dense, decoder-only language models available in various sizes, with some variants utilizing Mixture-of-Experts. These models are pretrained on large-scale datasets, supporting extended context lengths and multilingual communication. The family includes specialized models for coding, mathematics, and multimodal tasks, such as vision and audio processing.