Qwen 3.5 397B-A17B — B200 vs H100
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and H100 (NVIDIA Hopper) on Qwen 3.5 397B-A17B. 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.
At 64 tok/s/user interactivity on Qwen 3.5 397B-A17B, B200 delivers 3271 tok/s/GPU at $0.17 per million tokens; H100 delivers 480 tok/s/GPU at $0.77. B200 is 363% cheaper per token; B200 delivers 582% more tok/s/GPU at this point.
B200 posts 1805 tok/s/GPU for $0.29 per million tokens at 98 tok/s/user on Qwen 3.5 397B-A17B; H100 posts 219 tok/s/GPU for $1.67. B200 is 468% cheaper per token; B200 delivers 725% more tok/s/GPU.
Throughput at 132 tok/s/user on Qwen 3.5 397B-A17B: B200 hits 1074 tok/s/GPU, H100 hits 85. Per-million costs land at $0.50 and $4.12 respectively. B200 is 718% cheaper per token; B200 delivers 1168% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:3271.4H100:479.6 | B200:1804.9H100:218.9 | B200:1074.3H100:84.7 |
| Cost ($/M tok) | B200:$0.166H100:$0.768 | B200:$0.295H100:$1.672 | B200:$0.504H100:$4.121 |
| tok/s/MW | B200:1507578H100:277201 | B200:831767H100:126508 | B200:495059H100:48976 |
| Concurrency | B200:~110H100:~35 | B200:~41H100:~10 | B200:~18H100:~3 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.