DeepSeek R1 — H100 vs Vera Rubin NVL72
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on DeepSeek R1. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
H100: 739 tok/s/chip, $0.44 per million tokens at 45 tok/s/user on DeepSeek R1. Vera Rubin NVL72 is unmeasured here.
At 71 tok/s/user on DeepSeek R1, H100 delivers 281 tok/s/chip at $1.16 per million tokens; Vera Rubin NVL72 hasn't been benchmarked at this target.
H100 hits 154 tok/s/chip for $2.11 per million tokens at 97 tok/s/user on DeepSeek R1. No Vera Rubin NVL72 data at this operating point. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:739.5Vera Rubin NVL72:— | H100:280.6Vera Rubin NVL72:— | H100:154.3Vera Rubin NVL72:— |
| Cost ($/M tok) | H100:$0.439Vera Rubin NVL72:— | H100:$1.158Vera Rubin NVL72:— | H100:$2.106Vera Rubin NVL72:— |
| tok/s/MW | H100:539778Vera Rubin NVL72:— | H100:204852Vera Rubin NVL72:— | H100:112644Vera Rubin NVL72:— |
| Concurrency | H100:~132Vera Rubin NVL72:— | H100:~41Vera Rubin NVL72:— | H100:~17Vera Rubin NVL72:— |