Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on Muse Spark Open 816B. 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
Near the low end of the 84β389 tok/s/user interactivity band, at 160 tok/s/user on Muse Spark Open 816B: B200 runs 12485 tok/s/chip at $0.04/M tokens, GB300 NVL72 runs 16625 at $0.04/M. Cost per token is essentially tied; GB300 NVL72 delivers 33% more tok/s/chip.
Setting 236 tok/s/user as the target on Muse Spark Open 816B, B200 produces 6167 tok/s/chip ($0.08 per million tokens) and GB300 NVL72 produces 7661 ($0.08). B200 is 7% cheaper per token; GB300 NVL72 delivers 24% more tok/s/chip.
At 313 tok/s/user interactivity on Muse Spark Open 816B, B200 delivers 3827 tok/s/chip at $0.13 per million tokens; GB300 NVL72 delivers 5367 tok/s/chip at $0.12. GB300 NVL72 is 5% cheaper per token; GB300 NVL72 delivers 40% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:12484.6GB300 NVL72:16624.8 | B200:6166.9GB300 NVL72:7660.6 | B200:3826.6GB300 NVL72:5367.2 |
| Cost ($/M tok) | B200:$0.038GB300 NVL72:$0.039 | B200:$0.078GB300 NVL72:$0.084 | B200:$0.126GB300 NVL72:$0.120 |
| tok/s/MW | B200:7300916GB300 NVL72:7841868 | B200:3606381GB300 NVL72:3613477 | B200:2237771GB300 NVL72:2531709 |
| Concurrency | B200:~17GB300 NVL72:~12 | B200:~9GB300 NVL72:~5 | B200:~5GB300 NVL72:~3 |
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.
