GLM 5/5.1 — B300 vs GB200 NVL72
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on GLM 5/5.1. 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.
Near the low end of the 30–147 tok/s/user interactivity band, at 59 tok/s/user on GLM 5/5.1: B300 runs 2598 tok/s/GPU at $0.25/M tokens, GB200 NVL72 runs 7372 at $0.08/M. GB200 NVL72 is 199% cheaper per token; GB200 NVL72 delivers 184% more tok/s/GPU.
Setting 88 tok/s/user as the target on GLM 5/5.1, B300 produces 1755 tok/s/GPU ($0.37 per million tokens) and GB200 NVL72 produces 2207 ($0.28). GB200 NVL72 is 34% cheaper per token; GB200 NVL72 delivers 26% more tok/s/GPU.
At 118 tok/s/user interactivity on GLM 5/5.1, B300 delivers 1278 tok/s/GPU at $0.50 per million tokens; GB200 NVL72 delivers 1157 tok/s/GPU at $0.52. B300 is 3% cheaper per token; B300 delivers 10% more tok/s/GPU at this point. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B300:2597.9GB200 NVL72:7372.3 | B300:1754.8GB200 NVL72:2207.0 | B300:1277.6GB200 NVL72:1157.0 |
| Cost ($/M tok) | B300:$0.252GB200 NVL72:$0.084 | B300:$0.370GB200 NVL72:$0.276 | B300:$0.505GB200 NVL72:$0.522 |
| tok/s/MW | B300:1367336GB200 NVL72:3942382 | B300:923579GB200 NVL72:1180215 | B300:672435GB200 NVL72:618715 |
| Concurrency | B300:~20GB200 NVL72:~1318 | B300:~9GB200 NVL72:~85 | B300:~5GB200 NVL72:~73 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.