Models
Choose your model.
From development to production to on-robot deployment — same architecture, different constraints.
ClarkenAI Spark
7B
Any hardware. Zero friction.
Runs on any CPU. Built for developers who need to move fast — prototype, test, and integrate without touching production hardware.
Parameters7B
PrecisionINT8
HardwareAny CPU, 8 GB+
Latency~50ms
GPUNot required
CloudNot required
Ideal for
- —Development & prototyping
- —CI pipeline inference
- —Integration testing
Free
Download SDKClarkenAI Core
70B
Full reasoning. Your infrastructure.
Maximum intelligence for production workloads. Deploy on-premises or in the cloud — no constraints on where it runs.
Parameters70B
PrecisionINT8 / INT4
HardwareServer / Cloud
LatencyCloud-dependent
GPUOptional
CloudOptional
Ideal for
- —Cloud production deployments
- —On-prem server inference
- —Maximum capability workloads
Contact us
Contact usClarkenAI Edge
70B
70B. On your robot. Under 2ms.
Hybrid Matrix Architecture runs 70B parameter reasoning on a single CPU node. No GPU. No internet. No compromise. This is Ankatos Cortex v1.
Parameters70B
ArchitectureINT8 · 1-bit · INT8
HardwareSingle CPU node
Latency< 2ms target
GPUNot required
CloudNot required
Ideal for
- —Autonomous robots
- —Air-gapped environments
- —Real-time on-device decisions
Contact us
Get early accessFull comparison
| Spec | Spark | Core | Edge |
|---|---|---|---|
| Parameters | 7B | 70B | 70B |
| Architecture | INT8 | INT8 / INT4 | INT8·1-bit·INT8 |
| Latency | ~50ms | Cloud-dep. | < 2ms target |
| GPU required | No | Optional | No |
| Cloud required | No | Optional | No |
| Min. RAM | 8 GB | 80+ GB | 128 GB |
| Target | Dev / testing | Cloud prod | Robot on-device |
| Pricing | Free | Contact us | Contact us |