Vera Rubin NVL72 hits 5x GB300 per-GPU throughput on DeepSeek V4.1 Flash
ArtificialAnlys · x · 2026-11-07
NVIDIA submitted updated Vera Rubin NVL72 configs for DeepSeek V4.1 Flash to Artificial Analysis's AA-AgentPerf benchmark. Against the improved GB300 results:
- At 260 tok/s per agent, Vera Rubin delivers 5.0x GB300's throughput per GPU (669k vs 133k tok/s)
- At 80 tok/s per agent, 1.7x (805k vs 469k)
- Single agent: 1,249 tok/s vs 848 on GB300 (1.47x)
The advantage is largest in the high-speed regime, highlighting the new platform's throughput scaling for agentic, long-context workloads.
More from Infra
- Huawei Ascend 960 roadmap pulled forward up to 3 quarters; analyst says scale-up domain beats efficiency — teortaxesTex · 2026-11-07
- CroissAttn runs all six attention matmuls in FP4 on Blackwell: 1.5× throughput, 55% less memory — TristanThrush · 2026-11-07
- SemiAnalysis: SGLang on AMD Mi355 crushes B200/B300 in throughput test — AnushElangovan · 2026-11-07
- Musk explains Terafab: not 'terror fab' but a terawatt-scale fab — elonmusk · 2026-11-07
- AMD launches free AI Developer Program with cloud credits and ROCm training — AnushElangovan · 2026-11-07
- Musk says data center unpopularity is temporary: Memphis is booming with 2x wages — inductionheads · 2026-11-07