One device to power business AI across your teams.
Personal intelligence system that helps you write, express yourself, and get things done effortlessly.
One node, many models, adapters and agents
Meta, IBM, Nvidia, Mistral, DeepSeek, Qwen and many more
Future-proof and tailored to your organization, featuring multi-model capabilities and a wide range of domain-specific, fine-tuned adapters.
Compare Plans
Plans that adapt as your needs evolve
Base
499
Extended
999
Scale
1299
Enterprise
Custom
Lemony nodes
Users
AI models
Storage
AI model updates
Support Plan Update
Tokens/sec per user (Llama 4 17B)
GB RAM (273 GB/s bandwidth)
ARMv9 cores
TB NVMe storage
Max power (USB-C)
TOPS (FP8)
TOPS (FP32)
Tokens/year (Llama 4 17B)
Stack-to-Scale
Max model size
1x
5 Users / 1 Team
1
1TB
~25
128
14
4
90W
Up to 1,000
Up to 7.8
Up to 4B
48 GB/s (12-lane PCLe)
90 billion parameters
2x
.avif)
25 Users / 5 Teams
3
4TB
50$/mth/node
49$/mth direct support
~32
256
28
8
180W
Up to 2,000
Up to 15.6
Up to 26B
48 GB/s (12-lane PCLe)
120 billion parameters
4x
50 Users / 10 Team
6
8TB
50$/mth/node
49$/mth direct support
~40
384
42
12
270W
Up to 3,000
Up to 23.4
Up to 46B
48 GB/s (12-lane PCLe)
140 billion parameters
Up to 500 users
8
48TB
~60
512
56
16
360W
Up to 4,000
Up to 31.2
Up to 95B
48 GB/s (12-lane PCLe)
180 billion parameters
Lemony Specs
Lemony Node
Power
max. 80W
USB Type C
110/230V Power Adapter (included)
Data
Direct Connect: USB-C Adapter (included)
Multi User: RJ45 Ethernet (included)
Cluster: RJ45 10Gbit/s (included)
Supported AI Models
Open-Source LLMs
Multimodal LLMs
Custom Models
Small Language Models
Default Preloaded
Llama 3.2 11B
Llama 3.1 8b
Llama 3.2 1B
AI Unit
NPU
AI Accelerator Cluster
Size
9.45” x 8.66” x 3.74” 240 x
220 x 95mm
Weight
1.68 lbs
760 g
Still have questions?
.svg)
