Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on GLM 5.3 744B. 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 →
Throughput at 143 tok/s/user on GLM 5.3 744B: B200 hits 10203 tok/s/chip, GB200 NVL72 hits 12412. Per-million costs land at $0.05 and $0.04 respectively. GB200 NVL72 is 13% cheaper per token; GB200 NVL72 delivers 22% more tok/s/chip.
B200 / GB200 NVL72 on GLM 5.3 744B at 195 tok/s/user: 7040 / 8529 tok/s/chip, $0.07 / $0.06 per million tokens. GB200 NVL72 is 13% cheaper per token; GB200 NVL72 delivers 21% more tok/s/chip.
Toward the upper edge of the 92–298 tok/s/user interactivity band, at 247 tok/s/user on GLM 5.3 744B: B200 runs 4408 tok/s/chip at $0.11/M tokens, GB200 NVL72 runs 4946 at $0.10/M. GB200 NVL72 is 4% cheaper per token; GB200 NVL72 delivers 12% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:10203.2GB200 NVL72:12412.1 | B200:7039.8GB200 NVL72:8528.6 | B200:4408.0GB200 NVL72:4946.4 |
| Cost ($/M tok) | B200:$0.047GB200 NVL72:$0.042 | B200:$0.068GB200 NVL72:$0.061 | B200:$0.109GB200 NVL72:$0.104 |
| tok/s/MW | B200:5966757GB200 NVL72:6637513 | B200:4116854GB200 NVL72:4560734 | B200:2577796GB200 NVL72:2645149 |
| Concurrency | B200:~37GB200 NVL72:~63 | B200:~8GB200 NVL72:~39 | B200:~5GB200 NVL72:~5 |
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.
