Parameters
600M
Context Length
16K
Modality
Text
Architecture
Dense
KV Cache Rate
32 KB/tok
License
other
Release Date
5 Oct 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Self-hosted only
Precision
Context
Workload Memory: 3.03 GB
INT4 · 8K context
Base Weights: 0.28 GB
KV Cache: 0.25 GB
Runtime & Graphs: 2.50 GB
1x NVIDIA L4 24GB
3.0 / 21.6 GB usable
1x RTX 3060 12GB
3.0 / 10.8 GB usable
Apple M4 24GB
3.0 / 18.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Checkpoint format
FP16
CUDA
>= 11.8
PyTorch
>= 2.1
Transformers
>= 4.40.1
| Architecture | Precision | Attention kernel |
|---|---|---|
Blackwell, Hopper | Native FP16 | FlashAttention-3 |
Ada Lovelace, Ampere | Native FP16 | FlashAttention-2 |
No evaluation benchmarks for Liquid D1-Omni 600M available.
Overall Rank
-
Coding Rank
-
Liquid D1-Omni 600M is a compact multimodal omni model developed by Liquid AI designed for efficient edge deployment across audio, text, and visual inputs. It leverages liquid neural network architectures for real-time streaming and on-device processing.
Attention
Attention Structure
Grouped-Query Attention
Attention Heads
16
Key-Value Heads
8
Attention Head Dimension
64
Position Embedding
ROPE
RoPE Theta
1,000,000
Linear Attention
Yes
Linear Attention Ratio
62.5%
Normalization
RMS Normalization
Activation Function
SwigLU
Dimensions
Base Precision
FP16
Hidden Dimension Size
1,024
Number of Layers
16
FFN Intermediate Size (Dense)
6,656
Tokenizer
Vocabulary Size
65,536
The LFM2 model family developed by Liquid AI.