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 40–84 tok/s/user interactivity band, at 51 tok/s/user on GLM 5/5.1: B300 runs 2944 tok/s/GPU at $0.22/M tokens, GB200 NVL72 runs 1781 at $0.35/M. B300 is 56% cheaper per token; B300 delivers 65% more tok/s/GPU.
Setting 62 tok/s/user as the target on GLM 5/5.1, B300 produces 2476 tok/s/GPU ($0.26 per million tokens) and GB200 NVL72 produces 1299 ($0.47). B300 is 80% cheaper per token; B300 delivers 91% more tok/s/GPU.
At 73 tok/s/user interactivity on GLM 5/5.1, B300 delivers 2105 tok/s/GPU at $0.31 per million tokens; GB200 NVL72 delivers 582 tok/s/GPU at $1.09. B300 is 254% cheaper per token; B300 delivers 262% more tok/s/GPU at this point. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B300:2944.1GB200 NVL72:1781.1 | B300:2475.7GB200 NVL72:1299.1 | B300:2104.8GB200 NVL72:581.5 |
| Cost ($/M tok) | B300:$0.222GB200 NVL72:$0.345 | B300:$0.264GB200 NVL72:$0.474 | B300:$0.307GB200 NVL72:$1.089 |
| tok/s/MW | B300:1356714GB200 NVL72:848123 | B300:1140882GB200 NVL72:618597 | B300:969961GB200 NVL72:276919 |
| Concurrency | B300:~26GB200 NVL72:~145 | B300:~19GB200 NVL72:~130 | B300:~13GB200 NVL72:~109 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.