DeepSeek R1 — GB200 NVL72 vs MI325X Performance per Dollar
Cost per million tokens of GB200 NVL72 (NVIDIA Blackwell) versus MI325X (AMD CDNA 3) on DeepSeek R1. 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 DeepSeek R1 at 37 tok/s/user, the per-million math comes out to $0.07 for GB200 NVL72 and $0.47 for MI325X; GB200 NVL72 delivers 543% more output per dollar.
At 43 tok/s/user on DeepSeek R1, GB200 NVL72 costs $0.07 per million tokens; MI325X costs $0.60. GB200 NVL72 is 709% more cost-efficient at this operating point.
GB200 NVL72 edges MI325X at 48 tok/s/user on DeepSeek R1 — $0.08 per million tokens versus $0.83, a 1005% cost-per-token gap. (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): GB200 NVL72 $1.86/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 | GB200 NVL72:$0.073MI325X:$0.470 | GB200 NVL72:$0.074MI325X:$0.598 | GB200 NVL72:$0.075MI325X:$0.832 |
| Concurrency | GB200 NVL72:~2484MI325X:~17 | GB200 NVL72:~4883MI325X:~11 | GB200 NVL72:~6142MI325X:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.