The new Drex Max v1 model runs on Apple Silicon Macs and NVIDIA GPUs, giving enterprises a high‑capacity option for local routing and policy‑check workloads.
Nace AI has unveiled Drex Max v1, a 27‑billion‑parameter decision engine designed for on‑premise inference. The model runs natively on Apple Silicon Macs and NVIDIA GPUs, offering enterprises a high‑capacity solution for local routing and policy‑check workloads without relying on cloud services.
Key Features of Drex Max v1
The engine combines a massive parameter count with optimisations for both Apple Silicon and CUDA‑enabled GPUs, delivering low‑latency inference for complex decision‑making tasks. Nace AI highlights its ability to handle real‑time policy enforcement, fraud detection, and dynamic routing in data‑center environments.
Hardware Compatibility
Drex Max v1 leverages the unified memory architecture of Apple Silicon, allowing developers to run the model directly on MacBook Pro and Mac Studio machines. On the NVIDIA side, the engine supports the latest Ampere and Hopper GPUs, scaling performance across multiple cards for larger workloads.
Enterprise Use Cases
Enterprises can deploy Drex Max v1 for on‑premise policy checks that require immediate response times, such as network traffic routing, compliance verification, and real‑time recommendation engines. By keeping inference local, organisations reduce data‑transfer costs and mitigate latency spikes associated with cloud‑based AI services.
- Local routing and traffic management
- Policy‑check and compliance enforcement
- Fraud detection in financial transactions
- Real‑time recommendation systems
Nace AI also provides a suite of tools for model quantisation and pruning, enabling customers to tailor the engine’s footprint to their specific hardware constraints while preserving accuracy.
For readers interested in broader AI developments, explore the AI coverage on uxc.news for related stories and analysis.
Source: MarkTechPost coverage of Nace AI launches Drex Max v1