K2 Horizon 7B
Parameters
7B
Context Length
524K
Modality
Text
Architecture
Dense
KV Cache Rate
144 KB/tok
License
Apache 2.0
Release Date
1 Sept 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
Hardware Sizing
Precision
Context
Workload Memory: 6.88 GB
INT4 · 8K context
Base Weights: 3.26 GB
KV Cache: 1.13 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
6.9 / 21.6 GB usable
1x RTX 3060 12GB
6.9 / 10.8 GB usable
Apple M4 24GB
6.9 / 18.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Architecture Diagram
Evaluation Benchmarks
Rank
#124
| Benchmark | Score | Rank |
|---|---|---|
Agentic Index | 0.26 | 67 |
Coding Index | 0.39 | 109 |
Intelligence Index | 0.21 | 157 |
Rankings
Overall Rank
#124
Coding Rank
#104
About K2 Horizon 7B
K2 Horizon 7B is a dense 7-billion parameter language model offering robust instruction following and analytical capabilities. It serves as an accessible base for developers requiring efficient reasoning and domain fine-tuning.
Technical Specifications
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
32
Key-Value Heads
8
Attention Head Dimension
128
Position Embedding
ROPE
RoPE Theta
10,000,000
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Hidden Dimension Size
4,096
Number of Layers
36
FFN Intermediate Size (Dense)
12,288
Tokenizer
Vocabulary Size
250,624
Resources
About Kimi K2
Moonshot AI's Kimi K2 is a Mixture-of-Experts model featuring one trillion total parameters, activating 32 billion per token. Designed for agentic intelligence, it utilizes a sparse architecture with 384 experts and the MuonClip optimizer for training stability, supporting a 128K token context window.