Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on GLM 5.3 744B. 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
B200 / Vera Rubin NVL72 on GLM 5.3 744B at 165 tok/s/user: 11996 / 79081 tok/s/chip, $0.04 / $0.01 per million tokens. Vera Rubin NVL72 is 216% cheaper per token; Vera Rubin NVL72 delivers 559% more tok/s/chip.
Around the middle of the 112β324 tok/s/user interactivity band, at 218 tok/s/user on GLM 5.3 744B: B200 runs 8337 tok/s/chip at $0.06/M tokens, Vera Rubin NVL72 runs 47706 at $0.02/M. Vera Rubin NVL72 is 174% cheaper per token; Vera Rubin NVL72 delivers 472% more tok/s/chip.
Setting 271 tok/s/user as the target on GLM 5.3 744B, B200 produces 6542 tok/s/chip ($0.07 per million tokens) and Vera Rubin NVL72 produces 27842 ($0.04). Vera Rubin NVL72 is 104% cheaper per token; Vera Rubin NVL72 delivers 326% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:11996.1Vera Rubin NVL72:79080.7 | B200:8336.7Vera Rubin NVL72:47705.9 | B200:6542.2Vera Rubin NVL72:27842.1 |
| Cost ($/M tok) | B200:$0.040Vera Rubin NVL72:$0.013 | B200:$0.058Vera Rubin NVL72:$0.021 | B200:$0.073Vera Rubin NVL72:$0.036 |
| tok/s/MW | B200:7015247Vera Rubin NVL72:23963852 | B200:4875250Vera Rubin NVL72:14456334 | B200:3825851Vera Rubin NVL72:8436988 |
| Concurrency | B200:~8Vera Rubin NVL72:~274 | B200:~4Vera Rubin NVL72:~134 | B200:~2Vera Rubin NVL72:~57 |
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.
