Qwen 3.5 397B-A17B — H100 vs H200 Performance per Dollar
Cost per million tokens of H100 (NVIDIA Hopper) versus H200 (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 64 tok/s/user, the per-million math comes out to $0.85 for H100 and $0.77 for H200; H200 delivers 10% more output per dollar.
At 100 tok/s/user on Qwen 3.5 397B-A17B, H100 costs $1.15 per million tokens; H200 costs $1.07. H200 is 8% more cost-efficient at this operating point.
H200 edges H100 at 135 tok/s/user on Qwen 3.5 397B-A17B — $1.27 per million tokens versus $1.54, a 21% 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): H100 $1.30/GPU/hr · H200 $1.41/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 | H100:$0.848H200:$0.767 | H100:$1.151H200:$1.065 | H100:$1.536H200:$1.273 |
| Concurrency | H100:~28H200:~32 | H100:~13H200:~15 | H100:~7H200:~9 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.