Falcon3-3B
Parameters
3B
Context Length
33K
Modality
Text
Architecture
Dense
KV Cache Rate
168 KB/tok
License
TII Falcon-LLM License 2.0
Release Date
17 Dec 2024
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
Hardware & System Requirements
Precision
Context
Workload Memory: 5.21 GB
INT4 · 8K context
Base Weights: 1.40 GB
KV Cache: 1.31 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
5.2 / 21.6 GB usable
1x RTX 3060 12GB
5.2 / 10.8 GB usable
Apple M4 24GB
5.2 / 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
>= 4.46.1
| Architecture | Precision | Attention kernel |
|---|---|---|
Blackwell, Hopper | Native FP16 | FlashAttention-3 |
Ada Lovelace, Ampere | Native FP16 | FlashAttention-2 |
Architecture Diagram
Evaluation Benchmarks
No evaluation benchmarks for Falcon3-3B available.
Rankings
Overall Rank
-
Coding Rank
-
About Falcon3-3B
Falcon3-3B is a 3B parameter open model developed by TII balancing efficient inference with strong mathematical and instruction-following capabilities. It supports context lengths up to 32K tokens and quantization down to 1.58 Bitnet.
Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
24
Key-Value Heads
6
Attention Head Dimension
256
Position Embedding
ROPE
RoPE Theta
1,000,042
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Base Precision
FP16
Hidden Dimension Size
1,536
Number of Layers
28
FFN Intermediate Size (Dense)
9,216
Tokenizer
Vocabulary Size
131,072
About Falcon 3
The TII Falcon 3 model family comprises open-source, decoder-only language models (1B-10B parameters) designed for efficiency. Key innovations include an extended 32K token context window, Grouped-Query Attention (GQA), and specialized versions for scientific and code-oriented applications. Some variants integrate Mamba-based architectures.