Head-to-head AI inference benchmark comparison of B300 (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
Throughput at 163 tok/s/user on GLM 5.3 744B: B300 hits 10978 tok/s/chip, Vera Rubin NVL72 hits 80413. Per-million costs land at $0.06 and $0.01 respectively. Vera Rubin NVL72 is 359% cheaper per token; Vera Rubin NVL72 delivers 632% more tok/s/chip.
B300 / Vera Rubin NVL72 on GLM 5.3 744B at 215 tok/s/user: 7000 / 49229 tok/s/chip, $0.09 / $0.02 per million tokens. Vera Rubin NVL72 is 340% cheaper per token; Vera Rubin NVL72 delivers 603% more tok/s/chip.
Toward the upper edge of the 112β319 tok/s/user interactivity band, at 267 tok/s/user on GLM 5.3 744B: B300 runs 4674 tok/s/chip at $0.13/M tokens, Vera Rubin NVL72 runs 28786 at $0.03/M. Vera Rubin NVL72 is 286% cheaper per token; Vera Rubin NVL72 delivers 516% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:10978.2Vera Rubin NVL72:80413.3 | B300:6999.9Vera Rubin NVL72:49229.4 | B300:4674.2Vera Rubin NVL72:28786.2 |
| Cost ($/M tok) | B300:$0.057Vera Rubin NVL72:$0.012 | B300:$0.090Vera Rubin NVL72:$0.020 | B300:$0.134Vera Rubin NVL72:$0.035 |
| tok/s/MW | B300:5778012Vera Rubin NVL72:24367655 | B300:3684151Vera Rubin NVL72:14918001 | B300:2460118Vera Rubin NVL72:8723102 |
| Concurrency | B300:~10Vera Rubin NVL72:~280 | B300:~7Vera Rubin NVL72:~141 | B300:~2Vera Rubin NVL72:~60 |
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.
