Qwen 3.5 397B-A17B — MI300X vs MI355X
Head-to-head AI inference benchmark comparison of MI300X (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.
MI300X / MI355X on Qwen 3.5 397B-A17B at 43 tok/s/user: 346 / 2808 tok/s/GPU, $0.89 / $0.15 per million tokens. MI355X is 510% cheaper per token; MI355X delivers 712% more tok/s/GPU.
Around the middle of the 34–71 tok/s/user interactivity band, at 53 tok/s/user on Qwen 3.5 397B-A17B: MI300X runs 190 tok/s/GPU at $1.63/M tokens, MI355X runs 2454 at $0.17/M. MI355X is 880% cheaper per token; MI355X delivers 1195% more tok/s/GPU.
Setting 62 tok/s/user as the target on Qwen 3.5 397B-A17B, MI300X produces 123 tok/s/GPU ($2.52 per million tokens) and MI355X produces 2189 ($0.19). MI355X is 1245% cheaper per token; MI355X delivers 1684% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MI300X:345.8MI355X:2808.3 | MI300X:189.5MI355X:2453.8 | MI300X:122.7MI355X:2189.4 |
| Cost ($/M tok) | MI300X:$0.893MI355X:$0.146 | MI300X:$1.633MI355X:$0.167 | MI300X:$2.517MI355X:$0.187 |
| tok/s/MW | MI300X:193162MI355X:1059737 | MI300X:105886MI355X:925957 | MI300X:68573MI355X:826203 |
| Concurrency | MI300X:~34MI355X:~132 | MI300X:~15MI355X:~90 | MI300X:~8MI355X:~71 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.