Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on Kimi K3 2.8T. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics β same interactions as the main inference chart.
AgentX replays real coding-agent sessions rather than fixed-length prompts, so context grows turn over turn and most of each request is served from cache instead of being recomputed. That turns the comparison into a systems question: KV transfer between nodes, prefix-aware routing, and cache capacity all move the curve alongside raw chip throughput. Fixed-sequence workloads stay the clean baseline for kernel and silicon performance, so the two scenarios answer different questions about the same hardware. Learn more about AgentX β
At 216 tok/s/user interactivity on Kimi K3 2.8T, H200 delivers 70 tok/s/chip at $4.85 per million tokens; Vera Rubin NVL72 delivers 25310 tok/s/chip at $0.04. Vera Rubin NVL72 is 12142% cheaper per token; Vera Rubin NVL72 delivers 36124% more tok/s/chip at this point.
H200 posts 70 tok/s/chip for $4.85 per million tokens at 432 tok/s/user on Kimi K3 2.8T; Vera Rubin NVL72 posts 6849 tok/s/chip for $0.15. Vera Rubin NVL72 is 3213% cheaper per token; Vera Rubin NVL72 delivers 9703% more tok/s/chip.
Throughput at 647 tok/s/user on Kimi K3 2.8T: H200 hits 70 tok/s/chip, Vera Rubin NVL72 hits 2424. Per-million costs land at $4.85 and $0.41 respectively. Vera Rubin NVL72 is 1072% cheaper per token; Vera Rubin NVL72 delivers 3369% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H200:69.9Vera Rubin NVL72:25309.9 | H200:69.9Vera Rubin NVL72:6849.5 | H200:69.9Vera Rubin NVL72:2423.6 |
| Cost ($/M tok) | H200:$4.850Vera Rubin NVL72:$0.040 | H200:$4.850Vera Rubin NVL72:$0.146 | H200:$4.850Vera Rubin NVL72:$0.414 |
| tok/s/MW | H200:51001Vera Rubin NVL72:7669674 | H200:51001Vera Rubin NVL72:2075591 | H200:51001Vera Rubin NVL72:734433 |
| Concurrency | H200:~1Vera Rubin NVL72:~95 | H200:~1Vera Rubin NVL72:~31 | H200:~1Vera Rubin NVL72:~6 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
