Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and B300 (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 84β334 tok/s/user interactivity band, at 147 tok/s/user on Muse Spark Open 816B: B200 runs 15572 tok/s/chip at $0.03/M tokens, B300 runs 17710 at $0.04/M. B200 is 15% cheaper per token; B300 delivers 14% more tok/s/chip.
Setting 209 tok/s/user as the target on Muse Spark Open 816B, B200 produces 7243 tok/s/chip ($0.07 per million tokens) and B300 produces 10645 ($0.06). B300 is 13% cheaper per token; B300 delivers 47% more tok/s/chip.
At 272 tok/s/user interactivity on Muse Spark Open 816B, B200 delivers 4930 tok/s/chip at $0.10 per million tokens; B300 delivers 5819 tok/s/chip at $0.11. B200 is 11% cheaper per token; B300 delivers 18% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:15571.9B300:17710.0 | B200:7242.8B300:10645.2 | B200:4930.4B300:5819.2 |
| Cost ($/M tok) | B200:$0.031B300:$0.035 | B200:$0.066B300:$0.059 | B200:$0.097B300:$0.108 |
| tok/s/MW | B200:9106401B300:9321038 | B200:4235568B300:5602760 | B200:2883294B300:3062727 |
| Concurrency | B200:~21B300:~14 | B200:~10B300:~8 | B200:~7B300:~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.
