Qwen 3.5 397B-A17B — B200 vs H100 Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus H100 (NVIDIA Hopper) on Qwen 3.5 397B-A17B. Owning-hyperscaler TCO normalized by output tokens — performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
On Qwen 3.5 397B-A17B at 66 tok/s/user, the per-million math comes out to $0.16 for B200 and $0.87 for H100; B200 delivers 446% more output per dollar.
At 101 tok/s/user on Qwen 3.5 397B-A17B, B200 costs $0.32 per million tokens; H100 costs $1.16. B200 is 266% more cost-efficient at this operating point.
B200 edges H100 at 136 tok/s/user on Qwen 3.5 397B-A17B — $0.49 per million tokens versus $1.55, a 214% cost-per-token gap. (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.)
GPU pricing (owning hyperscaler): B200 $1.95/GPU/hr · H100 $1.30/GPU/hr. Source: SemiAnalysis Market August 2025 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B200:$0.160H100:$0.873 | B200:$0.317H100:$1.159 | B200:$0.494H100:$1.553 |
| Concurrency | B200:~112H100:~26 | B200:~38H100:~12 | B200:~18H100:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.