BLAST RADIUS

Frontier Line and access · Benchmarks evals

Nvidia Reports Vera Rubin MLPerf Inference Results

Sep 16, 2026, first seen via NVIDIA agent runtimes & toolkits

Nvidia reported that Vera Rubin NVL72 preview submissions delivered up to 3.7 times GB300 throughput and substantially higher AgentX performance.

Read the original at blogs.nvidia.comOpens the publisher's site in a new tab

More in Frontier Line and access