Qwen 3.5 397B-A17B — GB300 NVL72 vs H200 Performance per Dollar
Cost per million tokens of GB300 NVL72 (NVIDIA Blackwell) 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.
GB300 NVL72: $0.56 per million tokens. H200: $0.81. Both at 69 tok/s/user on Qwen 3.5 397B-A17B, with GB300 NVL72 43% cheaper.
Around the middle of the 32–182 tok/s/user interactivity band — at 107 tok/s/user — GB300 NVL72 runs $1.35 per million tokens on Qwen 3.5 397B-A17B while H200 runs $1.11. H200 is the cheaper choice by 22%.
On Qwen 3.5 397B-A17B at 145 tok/s/user, the per-million math comes out to $2.57 for GB300 NVL72 and $1.39 for H200; H200 delivers 85% more output per dollar. (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): GB300 NVL72 $2.65/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 | GB300 NVL72:$0.563H200:$0.807 | GB300 NVL72:$1.351H200:$1.108 | GB300 NVL72:$2.568H200:$1.387 |
| Concurrency | GB300 NVL72:~113H200:~28 | GB300 NVL72:~21H200:~13 | GB300 NVL72:~9H200:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.