gpt-oss 120B — H100 vs MI325X Performance per Dollar
Cost per million tokens of H100 (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.
On gpt-oss 120B at 78 tok/s/user, the per-million math comes out to $0.08 for H100 and $0.22 for MI325X; H100 delivers 168% more output per dollar.
At 90 tok/s/user on gpt-oss 120B, H100 costs $0.09 per million tokens; MI325X costs $0.27. H100 is 196% more cost-efficient at this operating point.
H100 edges MI325X at 102 tok/s/user on gpt-oss 120B — $0.10 per million tokens versus $0.40, a 281% cost-per-token gap. (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.)
Chip pricing (owning hyperscaler): H100 $1.17/chip/hr · MI325X $1.10/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | H100:$0.084MI325X:$0.225 | H100:$0.092MI325X:$0.272 | H100:$0.104MI325X:$0.396 |
| Concurrency | H100:~64MI325X:~23 | H100:~64MI325X:~28 | H100:~64MI325X:~16 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.