NVIDIA Vera Rubin NVL72 tops MLPerf Inference v6.1 in debut

NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut

NVIDIA's Vera Rubin NVL72 topped MLPerf Inference v6.1 in its first run. It's the post-Blackwell flagship with 72 GPUs linked via NVLink, built for large-scale inference. The post doesn't disclose exact scores or comparison models—only claims "leading performance." For buyers, this suggests inference throughput and latency improvements over H100/B200, but detailed numbers are needed to calculate ROI.