Qwen 3.5 397B-A17B — B200 vs MI325X Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus MI325X (AMD CDNA 3) 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.
Near the low end of the 26–68 tok/s/user interactivity band — at 36 tok/s/user — B200 runs $0.04 per million tokens on Qwen 3.5 397B-A17B while MI325X runs $0.25. B200 is the cheaper choice by 471%.
On Qwen 3.5 397B-A17B at 47 tok/s/user, the per-million math comes out to $0.05 for B200 and $0.38 for MI325X; B200 delivers 696% more output per dollar.
At 57 tok/s/user on Qwen 3.5 397B-A17B, B200 costs $0.05 per million tokens; MI325X costs $0.57. B200 is 1002% more cost-efficient at this operating point. (Numbers reflect the default 8k/1k · fp8 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 · MI325X $1.10/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:$0.044MI325X:$0.250 | B200:$0.048MI325X:$0.382 | B200:$0.052MI325X:$0.573 |
| Concurrency | B200:~145MI325X:~32 | B200:~99MI325X:~16 | B200:~77MI325X:~9 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.