MiniMax M3 428B — H100 vs H200 Performance per Dollar
Cost per million tokens of H100 (NVIDIA Hopper) versus H200 (NVIDIA Hopper) 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.
Near the low end of the 15–260 tok/s/user interactivity band — at 76 tok/s/user — H100 runs $0.67 per million tokens on MiniMax M3 428B while H200 runs $0.55. H200 is the cheaper choice by 22%.
On MiniMax M3 428B at 138 tok/s/user, the per-million math comes out to $1.52 for H100 and $1.06 for H200; H200 delivers 43% more output per dollar.
At 199 tok/s/user on MiniMax M3 428B, H100 costs $2.31 per million tokens; H200 costs $1.73. H200 is 33% more cost-efficient at this operating point. (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): H100 $1.30/GPU/hr · H200 $1.41/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 | H100:$0.673H200:$0.551 | H100:$1.518H200:$1.062 | H100:$2.307H200:$1.731 |
| Concurrency | H100:~31H200:~20 | H100:~8H200:~5 | H100:~4H200:~4 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.