Llama 3.3 70B — H100 vs MI300X Performance per Dollar
Cost per million tokens of H100 (NVIDIA Hopper) versus MI300X (AMD CDNA 3) on Llama 3.3 70B. 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 35–109 tok/s/user interactivity band — at 53 tok/s/user — H100 runs $0.25 per million tokens on Llama 3.3 70B while MI300X runs $0.26. H100 is the cheaper choice by 3%.
On Llama 3.3 70B at 72 tok/s/user, the per-million math comes out to $0.45 for H100 and $0.46 for MI300X; H100 delivers 2% more output per dollar.
At 91 tok/s/user on Llama 3.3 70B, H100 costs $1.19 per million tokens; MI300X costs $0.77. MI300X is 55% 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 · 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 | H100:$0.249MI300X:$0.257 | H100:$0.454MI300X:$0.462 | H100:$1.194MI300X:$0.771 |
| Concurrency | H100:~64MI300X:~50 | H100:~50MI300X:~32 | H100:~14MI300X:~18 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.