MiniMax M3 428B — H200 vs MI355X Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus MI355X (AMD CDNA 4) 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.
Push MiniMax M3 428B to 18 tok/s/user and H200 lands at $0.27 per million tokens against MI355X's $0.18 — MI355X pulls ahead by 52%.
H200: $0.28 per million tokens. MI355X: $0.21. Both at 27 tok/s/user on MiniMax M3 428B, with MI355X 33% cheaper.
Toward the upper edge of the 9–46 tok/s/user interactivity band — at 37 tok/s/user — H200 runs $0.30 per million tokens on MiniMax M3 428B while MI355X runs $0.28. MI355X is the cheaper choice by 5%. (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): H200 $1.41/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 | H200:$0.269MI355X:$0.177 | H200:$0.278MI355X:$0.209 | H200:$0.297MI355X:$0.281 |
| Concurrency | H200:~184MI355X:~294 | H200:~111MI355X:~158 | H200:~77MI355X:~97 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.