NVIDIA Vera Rubin NVL72 Achieves up to 2x Blackwell Ultra Performance in MLPerf Training v6.1
<img width="768" height="432" src="https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-768x432.jpg" class="webfeedsFeaturedVisual wp-post-image" alt="" style="display: block; margin-bottom: 5px; clear:both;max-width: 100%;" link_thumbnail="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-768x432.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-179x101.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-300x169.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-625x352.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-645x363.jpg 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-660x370.jpg 660w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-500x281.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-160x90.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-362x204.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-196x110.jpg 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-1024x576.jpg 1024w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-960x540.jpg 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855.webp 1520w" sizes="auto, (max-width: 768px) 100vw, 768px" title="techblog_featured2_1520x855" />As AI agents tackle increasingly complex workflows, reducing time-to-train is critical for rapidly iterating on both models and the agents they power. MLPerf...<img width="768" height="432" src="https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-768x432.jpg" class="webfeedsFeaturedVisual wp-post-image" alt="" style="display: block; margin-bottom: 5px; clear:both;max-width: 100%;" link_thumbnail="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-768x432.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-179x101.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-300x169.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-625x352.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-645x363.jpg 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-660x370.jpg 660w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-500x281.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-160x90.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-362x204.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-196x110.jpg 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-1024x576.jpg 1024w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855-960x540.jpg 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/11/techblog_featured2_1520x855.webp 1520w" sizes="auto, (max-width: 768px) 100vw, 768px" title="techblog_featured2_1520x855" /><p>As AI agents tackle increasingly complex workflows, reducing time-to-train is critical for rapidly iterating on both models and the agents they power. MLPerf Training v6.1 marks the debut of the NVIDIA Vera Rubin architecture, and these early results demonstrate that its CUDA software stack is already highly mature. In its first appearance, the Vera Rubin NVL72 platform delivered the fastest time…</p>
<p><a href="https://developer.nvidia.com/blog/nvidia-vera-rubin-nvl72-achieves-up-to-2x-blackwell-ultra-performance-in-mlperf-training-v6-1/" rel="nofollow" data-wpel-link="internal" target="_self">Source</a></p>
Read original article ↗
Related Articles
NVIDIA DGX Spark Developer Kit Turns a Desk Into a Supercomputing Lab
At SC26, an annual supercomputing conference taking place this week in Chicago, NVIDIA is introducing the DGX Spark HPC
Molecular dynamics (MD) is an indispensable tool across physics, chemistry, biology, and materials science. Owing to its
Quantum Machines Delivers 1st Results from CUDA-Q QEC Real-Time Decoders
An NVIDIA CUDA-Q program compiled for the OPX1000 can now run real-time, mid-circuit decoding in microseconds on acceler
NVIDIA Announces Rubin HPC, a New GPU to Accelerate Agents for Science, Engineering and Analytics
A record leap in CUDA cores with accelerated data movement shorten time to results for science and engineering workloads