DeepSeek R1 · Performance per Dollar
DeepSeek R1 — GB300 NVL72 vs MI325X Performance per Dollar
Cost per million tokens of GB300 NVL72 (NVIDIA Blackwell) versus MI325X (AMD CDNA 3) on DeepSeek R1. Large-hyperscaler-volume ownership TCO normalized by total 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.
Push DeepSeek R1 to 40 tok/s/user and GB300 NVL72 lands at $0.073 per million tokens against MI325X's $0.34 — GB300 NVL72 costs 78% less.
GB300 NVL72: $0.077 per million tokens. MI325X: $0.42. Both at 60 tok/s/user on DeepSeek R1, with GB300 NVL72 82% cheaper.
Toward the upper edge of the 20–100 tok/s/user interactivity band — at 80 tok/s/user — GB300 NVL72 runs $0.100 per million tokens on DeepSeek R1 while MI325X runs $0.52. GB300 NVL72 is the cheaper choice by 81%. (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): GB300 NVL72 $2.31/chip/hr · MI325X $1.10/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollars per Million Tokens | GB300 NVL72$0.073MI325X$0.34 | GB300 NVL72$0.077MI325X$0.42 | GB300 NVL72$0.100MI325X$0.52 |
| Concurrency | GB300 NVL72~1,229MI325X~20 | GB300 NVL72~1,171MI325X~11 | GB300 NVL72~578MI325X~7 |
