Qwen 3.5 397B-A17B — MI300X vs MI325X
Head-to-head AI inference benchmark comparison of MI300X (AMD CDNA 3) 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.
At 55 tok/s/user interactivity on Qwen 3.5 397B-A17B, MI300X delivers 308 tok/s/GPU at $0.99 per million tokens; MI325X delivers 399 tok/s/GPU at $0.89. MI325X is 11% cheaper per token; MI325X delivers 30% more tok/s/GPU at this point.
MI300X posts 226 tok/s/GPU for $1.37 per million tokens at 68 tok/s/user on Qwen 3.5 397B-A17B; MI325X posts 285 tok/s/GPU for $1.25. MI325X is 10% cheaper per token; MI325X delivers 26% more tok/s/GPU.
Throughput at 82 tok/s/user on Qwen 3.5 397B-A17B: MI300X hits 151 tok/s/GPU, MI325X hits 185. Per-million costs land at $2.06 and $1.93 respectively. MI325X is 7% cheaper per token; MI325X delivers 22% more tok/s/GPU. (Numbers reflect the default 1k/1k · bf16 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MI300X:308.0MI325X:399.5 | MI300X:225.8MI325X:284.6 | MI300X:151.3MI325X:184.9 |
| Cost ($/M tok) | MI300X:$0.993MI325X:$0.891 | MI300X:$1.375MI325X:$1.252 | MI300X:$2.062MI325X:$1.926 |
| tok/s/MW | MI300X:172074MI325X:183247 | MI300X:126151MI325X:130561 | MI300X:84545MI325X:84815 |
| Concurrency | MI300X:~23MI325X:~32 | MI300X:~13MI325X:~17 | MI300X:~8MI325X:~10 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.