gpt-oss 120B — H200 vs MI325X Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus MI325X (AMD CDNA 3) on gpt-oss 120B. 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 40–113 tok/s/user interactivity band — at 58 tok/s/user — H200 runs $0.08 per million tokens on gpt-oss 120B while MI325X runs $0.14. H200 is the cheaper choice by 69%.
On gpt-oss 120B at 77 tok/s/user, the per-million math comes out to $0.09 for H200 and $0.25 for MI325X; H200 delivers 164% more output per dollar.
At 95 tok/s/user on gpt-oss 120B, H200 costs $0.11 per million tokens; MI325X costs $0.36. H200 is 225% more cost-efficient at this operating point. (Numbers reflect the default 1k/1k · fp4 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 · MI325X $1.28/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:$0.083MI325X:$0.139 | H200:$0.094MI325X:$0.249 | H200:$0.111MI325X:$0.360 |
| Concurrency | H200:~64MI325X:~64 | H200:~64MI325X:~29 | H200:~64MI325X:~23 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.