Llama 3.3 70B · Performance per Dollar

Llama 3.3 70B — H100 vs MI325X Performance per Dollar

Cost per million tokens of H100 (NVIDIA Hopper) versus MI325X (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.

H100: $0.25 per million tokens. MI325X: $0.24. Both at 53 tok/s/user on Llama 3.3 70B, with MI325X 3% cheaper.

Around the middle of the 35–109 tok/s/user interactivity band — at 72 tok/s/user — H100 runs $0.45 per million tokens on Llama 3.3 70B while MI325X runs $0.39. MI325X is the cheaper choice by 16%.

On Llama 3.3 70B at 91 tok/s/user, the per-million math comes out to $1.19 for H100 and $0.62 for MI325X; MI325X delivers 93% more output per dollar. (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 · MI325X $1.28/GPU/hr. Source: SemiAnalysis Market August 2025 Pricing Surveys & AI Cloud TCO Model.

View full latency + throughput comparison →

Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Dollar per Million Tokens
H100:$0.249MI325X:$0.243
H100:$0.454MI325X:$0.391
H100:$1.194MI325X:$0.618
Concurrency
H100:~64MI325X:~64
H100:~50MI325X:~59
H100:~14MI325X:~26

Inference Performance

Inference performance metrics across different models, hardware configurations, and serving parameters.