Qwen 3.5 397B-A17B — B200 vs GB300 NVL72 Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus GB300 NVL72 (NVIDIA Blackwell) 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.
B200: $0.05 per million tokens. GB300 NVL72: $0.02. Both at 98 tok/s/user on Qwen 3.5 397B-A17B, with GB300 NVL72 189% cheaper.
Around the middle of the 24–320 tok/s/user interactivity band — at 172 tok/s/user — B200 runs $0.09 per million tokens on Qwen 3.5 397B-A17B while GB300 NVL72 runs $0.04. GB300 NVL72 is the cheaper choice by 133%.
On Qwen 3.5 397B-A17B at 246 tok/s/user, the per-million math comes out to $0.13 for B200 and $0.06 for GB300 NVL72; GB300 NVL72 delivers 135% more output per dollar. (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.)
Chip pricing (owning hyperscaler): B200 $1.73/chip/hr · GB300 NVL72 $2.31/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B200:$0.053GB300 NVL72:$0.018 | B200:$0.089GB300 NVL72:$0.038 | B200:$0.133GB300 NVL72:$0.056 |
| Concurrency | B200:~96GB300 NVL72:~1181 | B200:~8GB300 NVL72:~393 | B200:~4GB300 NVL72:~145 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.