Qwen 3.5 397B-A17B — H100 vs MI325X
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI325X (AMD CDNA 3) on Qwen 3.5 397B-A17B. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
Throughput at 52 tok/s/user on Qwen 3.5 397B-A17B: H100 hits 308 tok/s/GPU, MI325X hits 273. Per-million costs land at $1.20 and $1.36 respectively. H100 is 14% cheaper per token; H100 delivers 13% more tok/s/GPU.
H100 / MI325X on Qwen 3.5 397B-A17B at 59 tok/s/user: 265 / 195 tok/s/GPU, $1.38 / $1.83 per million tokens. H100 is 33% cheaper per token; H100 delivers 36% more tok/s/GPU.
Toward the upper edge of the 45–72 tok/s/user interactivity band, at 66 tok/s/user on Qwen 3.5 397B-A17B: H100 runs 231 tok/s/GPU at $1.55/M tokens, MI325X runs 116 at $3.09/M. H100 is 100% cheaper per token; H100 delivers 99% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | H100:307.6MI325X:273.4 | H100:264.7MI325X:195.1 | H100:230.9MI325X:116.0 |
| Cost ($/M tok) | H100:$1.198MI325X:$1.363 | H100:$1.378MI325X:$1.833 | H100:$1.548MI325X:$3.094 |
| tok/s/MW | H100:177792MI325X:125411 | H100:153021MI325X:89501 | H100:133443MI325X:53202 |
| Concurrency | H100:~25MI325X:~22 | H100:~19MI325X:~14 | H100:~14MI325X:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.