Qwen 3.5 397B-A17B — GB200 NVL72 vs MI355X Performance per Dollar
Cost per million tokens of GB200 NVL72 (NVIDIA Blackwell) versus MI355X (AMD CDNA 4) 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.
GB200 NVL72 edges MI355X at 82 tok/s/user on Qwen 3.5 397B-A17B — $0.06 per million tokens versus $0.10, a 76% cost-per-token gap.
Push Qwen 3.5 397B-A17B to 129 tok/s/user and GB200 NVL72 lands at $0.13 per million tokens against MI355X's $0.15 — GB200 NVL72 pulls ahead by 13%.
GB200 NVL72: $0.24 per million tokens. MI355X: $0.24. Both at 176 tok/s/user on Qwen 3.5 397B-A17B, with GB200 NVL72 2% cheaper. (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.)
GPU pricing (owning hyperscaler): GB200 NVL72 $2.21/GPU/hr · MI355X $1.48/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 | GB200 NVL72:$0.059MI355X:$0.103 | GB200 NVL72:$0.133MI355X:$0.150 | GB200 NVL72:$0.238MI355X:$0.244 |
| Concurrency | GB200 NVL72:~1052MI355X:~22 | GB200 NVL72:~107MI355X:~9 | GB200 NVL72:~36MI355X:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.