NVIDIA Vera Rubin NVL72 leads MLPerf agentic inference performance
NVIDIA Vera Rubin NVL72 shows high agentic inference performance in MLPerf Endpoints v1.0 results.
- Provisional MLPerf Endpoints v1.0 results show NVIDIA Vera Rubin NVL72 delivers over 10x the throughput per kilowatt of NVIDIA GB300 NVL72.
- Across Kimi K3, DeepSeek-V4.1-Flash, and Qwen3.6-35B-A3B models, Vera Rubin NVL72 achieves 5x to 10x higher throughput per kilowatt.
- It also shows 3x to 5x higher end-to-end interactivity at matched concurrency compared with GB300 NVL72.
- Context-aware serving techniques like prefix caching reduce redundant computation and balance serving load.
Who it affects This matters to builders working with agentic AI workloads and inference systems.
What to do
Visit MLPerf Endpoints results to review the published benchmark data.
Source: developer.nvidia.com ·