Qwen 3.5 397B-A17B — MI325X vs MI355X
Head-to-head AI inference benchmark comparison of MI325X (AMD CDNA 3) and MI355X (AMD CDNA 4) 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.
Near the low end of the 37–72 tok/s/user interactivity band, at 46 tok/s/user on Qwen 3.5 397B-A17B: MI325X runs 403 tok/s/GPU at $0.86/M tokens, MI355X runs 2690 at $0.15/M. MI355X is 466% cheaper per token; MI355X delivers 568% more tok/s/GPU.
Setting 55 tok/s/user as the target on Qwen 3.5 397B-A17B, MI325X produces 219 tok/s/GPU ($1.63 per million tokens) and MI355X produces 2393 ($0.17). MI355X is 856% cheaper per token; MI355X delivers 991% more tok/s/GPU.
At 64 tok/s/user interactivity on Qwen 3.5 397B-A17B, MI325X delivers 134 tok/s/GPU at $2.64 per million tokens; MI355X delivers 2131 tok/s/GPU at $0.19. MI355X is 1270% cheaper per token; MI355X delivers 1487% more tok/s/GPU at this point. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MI325X:402.9MI355X:2690.3 | MI325X:219.4MI355X:2392.9 | MI325X:134.3MI355X:2131.3 |
| Cost ($/M tok) | MI325X:$0.864MI355X:$0.153 | MI325X:$1.631MI355X:$0.171 | MI325X:$2.638MI355X:$0.193 |
| tok/s/MW | MI325X:184830MI355X:1015220 | MI325X:100633MI355X:902983 | MI325X:61591MI355X:804277 |
| Concurrency | MI325X:~37MI355X:~116 | MI325X:~16MI355X:~85 | MI325X:~9MI355X:~68 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.