Tag

#mlperf

2 stories tagged mlperf.

Step chart showing NVIDIA Vera Rubin NVL72 throughput per megawatt progression on AgentX: H200 NVL8 at 1x baseline, GB300 NVL72 at 15x, and Vera Rubin NVL72 at 450x cumulative. The Vera Rubin NVL72 keyword marks the leading generation. Source: NVIDIA AgentX preview results.

NVIDIA Vera Rubin NVL72 debuts in MLPerf with 3.7x lead

NVIDIA Vera Rubin NVL72 enters MLPerf Inference v6.1 with 3.7x throughput over GB300 NVL72, plus 99% scaling efficiency and 30x gains on agentic workloads.

Lars Cornelissen · Sep 16, 2026

MLPerf Training 6.0 scale comparison showing Blackwell at 8,192 GPUs for DeepSeek-V3 and 8,192 GPUs for Llama 3.1 405B, with AMD OCI FLUX.1 at 512 GPUs.

MLPerf Training 6.0 gives Blackwell the scale edge

MLPerf Training 6.0 shows Blackwell leading all seven tests, but the useful signal is scale: 8,192 GPUs and MoE training pressure.

Lars Cornelissen · Jun 17, 2026