Inkling 2 Small
Active Parameters
2B
Context Length
1.05M
Modality
Text
Architecture
Mixture of Experts
KV Cache Rate
192 KB/tok
License
Apache 2.0
Release Date
4 Nov 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
Hardware & System Requirements
Precision
Context
Workload Memory: 12.85 GB
INT4 · 8K context
Base Weights: 8.85 GB
KV Cache: 1.50 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
12.8 / 21.6 GB usable
1x RTX 4060 Ti 16GB
12.8 / 14.4 GB usable
Apple M4 24GB
12.8 / 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
>= 4.40.1
| 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 Inkling 2 Small available.
Rankings
Overall Rank
-
Coding Rank
-
About Inkling 2 Small
Inkling 2 Small is an efficient multimodal language model built by Thinking Machine Labs. It offers fast inference and compact parameter size for on-device and edge deployment.
Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
48
Key-Value Heads
8
Attention Head Dimension
128
Position Embedding
ROPE
RoPE Theta
1,000,000
Sliding Window Attention
Yes
Sliding Window Size
1,024
Sliding Window Ratio
83.3%
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Base Precision
FP16
Hidden Dimension Size
4,608
Number of Layers
48
Multi-Token Prediction Heads
5
Tokenizer
Vocabulary Size
201,024
Mixture of Experts
Total Expert Parameters
17.0B
Number of Experts
384
Active Experts
8
Shared Experts
0
FFN Intermediate Size (per Expert)
2,304
Dense Layers Before MoE
0
Resources
About Inkling
Inkling is a family of open-weights multimodal Mixture-of-Experts (MoE) models developed by Thinking Machine Labs.