NVIDIA releases Dynamo deployment recipes for Korean open models

NVIDIA is releasing Dynamo-optimized deployment recipes for Korean open models, including A.X K2 from SK Telecom, K-EXAONE 2.0 from LG AI Research, Solar Open 2 from Upstage and Motif-3 from Motif Technologies. The recipes are designed to help developers bring these models into production faster and make AI services more broadly accessible.

Models that exceed the capacity of a single GPU must be distributed across multiple GPUs. Maintaining performance requires careful coordination of request routing and communication. NVIDIA Dynamo, an open source distributed inference-serving framework, handles those tasks for models deployed in multi-node environments at data center scale.

The deployment recipes are optimized for NVIDIA B200 GPUs and deliver up to 2x performance out of the box, according to NVIDIA. They aim to reduce deployment guesswork while supporting the infrastructure needed to serve large models.