Muse Spark Open 816B — B200 vs Vera Rubin NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on Muse Spark Open 816B. 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.
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 →
At 193 tok/s/user interactivity on Muse Spark Open 816B, B200 delivers 17909 tok/s/chip at $0.03 per million tokens; Vera Rubin NVL72 delivers 85038 tok/s/chip at $0.01. Vera Rubin NVL72 is 128% cheaper per token; Vera Rubin NVL72 delivers 375% more tok/s/chip at this point.
B200 posts 10515 tok/s/chip for $0.05 per million tokens at 280 tok/s/user on Muse Spark Open 816B; Vera Rubin NVL72 posts 60397 tok/s/chip for $0.02. Vera Rubin NVL72 is 175% cheaper per token; Vera Rubin NVL72 delivers 474% more tok/s/chip.
Throughput at 366 tok/s/user on Muse Spark Open 816B: B200 hits 4364 tok/s/chip, Vera Rubin NVL72 hits 47891. Per-million costs land at $0.11 and $0.02 respectively. Vera Rubin NVL72 is 426% cheaper per token; Vera Rubin NVL72 delivers 997% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:17908.7Vera Rubin NVL72:85037.6 | B200:10515.2Vera Rubin NVL72:60397.1 | B200:4363.7Vera Rubin NVL72:47891.3 |
| Cost ($/M tok) | B200:$0.027Vera Rubin NVL72:$0.012 | B200:$0.046Vera Rubin NVL72:$0.017 | B200:$0.110Vera Rubin NVL72:$0.021 |
| tok/s/MW | B200:10472910Vera Rubin NVL72:25768957 | B200:6149236Vera Rubin NVL72:18302137 | B200:2551855Vera Rubin NVL72:14512506 |
| Concurrency | B200:~11Vera Rubin NVL72:~48 | B200:~10Vera Rubin NVL72:~32 | B200:~4Vera Rubin NVL72:~24 |