Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 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
Throughput at 123 tok/s/user on Muse Spark Open 816B: B300 hits 20311 tok/s/chip, GB200 NVL72 hits 18284. Per-million costs land at $0.03 and $0.03 respectively. GB200 NVL72 is 9% cheaper per token; B300 delivers 11% more tok/s/chip.
B300 / GB200 NVL72 on Muse Spark Open 816B at 189 tok/s/user: 13005 / 8872 tok/s/chip, $0.05 / $0.06 per million tokens. B300 is 21% cheaper per token; B300 delivers 47% more tok/s/chip.
Toward the upper edge of the 57β321 tok/s/user interactivity band, at 255 tok/s/user on Muse Spark Open 816B: B300 runs 6048 tok/s/chip at $0.10/M tokens, GB200 NVL72 runs 5701 at $0.09/M. GB200 NVL72 is 15% cheaper per token; B300 delivers 6% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:20310.7GB200 NVL72:18283.8 | B300:13004.6GB200 NVL72:8872.1 | B300:6047.9GB200 NVL72:5700.9 |
| Cost ($/M tok) | B300:$0.031GB200 NVL72:$0.028 | B300:$0.048GB200 NVL72:$0.058 | B300:$0.104GB200 NVL72:$0.091 |
| tok/s/MW | B300:10689865GB200 NVL72:9777452 | B300:6844537GB200 NVL72:4744416 | B300:3183094GB200 NVL72:3048613 |
| Concurrency | B300:~15GB200 NVL72:~13 | B300:~10GB200 NVL72:~7 | B300:~4GB200 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.
