Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on Kimi K3 2.8T. 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 →
Near the low end of the 9–205 tok/s/user interactivity band, at 58 tok/s/user on Kimi K3 2.8T: B200 runs 4037 tok/s/chip at $0.12/M tokens, GB300 NVL72 runs 11994 at $0.05/M. GB300 NVL72 is 122% cheaper per token; GB300 NVL72 delivers 197% more tok/s/chip.
Setting 107 tok/s/user as the target on Kimi K3 2.8T, B200 produces 2028 tok/s/chip ($0.24 per million tokens) and GB300 NVL72 produces 7712 ($0.08). GB300 NVL72 is 185% cheaper per token; GB300 NVL72 delivers 280% more tok/s/chip.
At 157 tok/s/user interactivity on Kimi K3 2.8T, B200 delivers 1252 tok/s/chip at $0.38 per million tokens; GB300 NVL72 delivers 5278 tok/s/chip at $0.12. GB300 NVL72 is 216% cheaper per token; GB300 NVL72 delivers 321% more tok/s/chip at this point. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:4037.5GB300 NVL72:11993.5 | B200:2028.4GB300 NVL72:7711.7 | B200:1252.5GB300 NVL72:5277.8 |
| Cost ($/M tok) | B200:$0.119GB300 NVL72:$0.054 | B200:$0.237GB300 NVL72:$0.083 | B200:$0.384GB300 NVL72:$0.122 |
| tok/s/MW | B200:2361083GB300 NVL72:5657330 | B200:1186200GB300 NVL72:3637613 | B200:732430GB300 NVL72:2489550 |
| Concurrency | B200:~14GB300 NVL72:~48 | B200:~5GB300 NVL72:~35 | B200:~2GB300 NVL72:~23 |
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.
