$NBIS earned $NVDA Exemplar Cloud validation for HGX B300 training after posting the fastest single node results in two MLPerf benchmarks.
HGX B300 performed within ~11% of GB300 NVL72 on those workloads giving Nebius more flexibility to deploy high performance compute without relying entirely on rack scale systems.
That is how Nebius moves toward delivering better performance and infrastructure efficiency for AI customers.
HGX B300 performed within ~11% of GB300 NVL72 on those workloads giving Nebius more flexibility to deploy high performance compute without relying entirely on rack scale systems.
That is how Nebius moves toward delivering better performance and infrastructure efficiency for AI customers.