Skip to content

NVIDIA Dynamo Recipes Bring Korea’s Leading Open Models to Production

NVIDIA AI

The recipes optimize multi-node inference for A.X K2, K-EXAONE 2.0, Solar Open 2, and Motif-3 on NVIDIA B200 GPUs.

Key points

  • Delivers up to 2x out-of-the-box inference performance gains on B200 hardware.
  • Supports SK Telecom's A.X K2, LG AI Research's K-EXAONE 2.0, Upstage's Solar Open 2, and Motif Technologies' Motif-3.
  • Uses NVIDIA Dynamo, an open-source distributed serving framework designed for multi-node data center routing and communication.

Why it matters

Developers serving these Korean open models on B200 infrastructure can deploy across multi-GPU setups without manual tuning for request routing and communication.

Sources