Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on DeepSeek V4.1 Flash 552B. 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 →
Setting 11 tok/s/user as the target on DeepSeek V4.1 Flash 552B, GB200 NVL72 produces 78100 tok/s/chip ($0.01 per million tokens) and GB300 NVL72 produces 17718 ($0.04). GB200 NVL72 is 447% cheaper per token; GB200 NVL72 delivers 341% more tok/s/chip.
At 15 tok/s/user interactivity on DeepSeek V4.1 Flash 552B, GB200 NVL72 delivers 78100 tok/s/chip at $0.01 per million tokens; GB300 NVL72 delivers 14268 tok/s/chip at $0.04. GB200 NVL72 is 580% cheaper per token; GB200 NVL72 delivers 447% more tok/s/chip at this point.
GB200 NVL72 posts 78100 tok/s/chip for $0.01 per million tokens at 18 tok/s/user on DeepSeek V4.1 Flash 552B; GB300 NVL72 posts 12449 tok/s/chip for $0.05. GB200 NVL72 is 679% cheaper per token; GB200 NVL72 delivers 527% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:78100.0GB300 NVL72:17718.5 | GB200 NVL72:78100.0GB300 NVL72:14268.1 | GB200 NVL72:78100.0GB300 NVL72:12449.2 |
| Cost ($/M tok) | GB200 NVL72:$0.007GB300 NVL72:$0.036 | GB200 NVL72:$0.007GB300 NVL72:$0.045 | GB200 NVL72:$0.007GB300 NVL72:$0.052 |
| tok/s/MW | GB200 NVL72:41764704GB300 NVL72:8357775 | GB200 NVL72:41764704GB300 NVL72:6730225 | GB200 NVL72:41764704GB300 NVL72:5872271 |
| Concurrency | GB200 NVL72:~64GB300 NVL72:~34 | GB200 NVL72:~64GB300 NVL72:~23 | GB200 NVL72:~64GB300 NVL72:~19 |
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.
