MiniMax M3 428B — GB300 NVL72 vs MI300X Performance per Dollar
Cost per million tokens of GB300 NVL72 (NVIDIA Blackwell) versus MI300X (AMD CDNA 3) on MiniMax M3 428B. 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.
GB300 NVL72: $0.28 per million tokens. MI300X: $0.59. Both at 47 tok/s/user on MiniMax M3 428B, with GB300 NVL72 110% cheaper.
Around the middle of the 13–149 tok/s/user interactivity band — at 81 tok/s/user — GB300 NVL72 runs $0.92 per million tokens on MiniMax M3 428B while MI300X runs $1.53. GB300 NVL72 is the cheaper choice by 67%.
On MiniMax M3 428B at 115 tok/s/user, the per-million math comes out to $2.34 for GB300 NVL72 and $2.14 for MI300X; MI300X delivers 9% more output per dollar. (Numbers reflect the default 1k/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): GB300 NVL72 $2.65/GPU/hr · MI300X $1.12/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 | GB300 NVL72:$0.280MI300X:$0.590 | GB300 NVL72:$0.921MI300X:$1.534 | GB300 NVL72:$2.342MI300X:$2.145 |
| Concurrency | GB300 NVL72:~649MI300X:~47 | GB300 NVL72:~96MI300X:~11 | GB300 NVL72:~29MI300X:~5 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.