Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on Qwen 3.8 Flash Next 176B-A6B. 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 64 tok/s/user on Qwen 3.8 Flash Next 176B-A6B: B300 hits 80846 tok/s/chip, Vera Rubin NVL72 hits 223506. Per-million costs land at $0.01 and $0.00 respectively. Vera Rubin NVL72 is 73% cheaper per token; Vera Rubin NVL72 delivers 176% more tok/s/chip.
B300 / Vera Rubin NVL72 on Qwen 3.8 Flash Next 176B-A6B at 104 tok/s/user: 68675 / 208116 tok/s/chip, $0.01 / $0.00 per million tokens. Vera Rubin NVL72 is 90% cheaper per token; Vera Rubin NVL72 delivers 203% more tok/s/chip.
Toward the upper edge of the 25β182 tok/s/user interactivity band, at 143 tok/s/user on Qwen 3.8 Flash Next 176B-A6B: B300 runs 49159 tok/s/chip at $0.01/M tokens, Vera Rubin NVL72 runs 177560 at $0.01/M. Vera Rubin NVL72 is 126% cheaper per token; Vera Rubin NVL72 delivers 261% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:80846.0Vera Rubin NVL72:223506.0 | B300:68674.6Vera Rubin NVL72:208115.7 | B300:49159.2Vera Rubin NVL72:177560.5 |
| Cost ($/M tok) | B300:$0.008Vera Rubin NVL72:$0.004 | B300:$0.009Vera Rubin NVL72:$0.005 | B300:$0.013Vera Rubin NVL72:$0.006 |
| tok/s/MW | B300:42550528Vera Rubin NVL72:67729102 | B300:36144527Vera Rubin NVL72:63065349 | B300:25873248Vera Rubin NVL72:53806202 |
| Concurrency | B300:~16Vera Rubin NVL72:~201 | B300:~14Vera Rubin NVL72:~272 | B300:~10Vera Rubin NVL72:~281 |
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.
