Qwen 3.5 397B-A17B — MI325X vs RTX PRO 6000
Head-to-head AI inference benchmark comparison of MI325X (AMD CDNA 3) and RTX PRO 6000 (NVIDIA Blackwell) 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.
MI325X: 399 tok/s/GPU, $0.89 per million tokens at 55 tok/s/user on Qwen 3.5 397B-A17B. RTX PRO 6000 is unmeasured here.
At 70 tok/s/user on Qwen 3.5 397B-A17B, MI325X delivers 274 tok/s/GPU at $1.30 per million tokens; RTX PRO 6000 hasn't been benchmarked at this target.
MI325X hits 170 tok/s/GPU for $2.09 per million tokens at 84 tok/s/user on Qwen 3.5 397B-A17B. No RTX PRO 6000 data at this operating point. (Numbers reflect the default 1k/1k · bf16 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MI325X:399.5RTX PRO 6000:— | MI325X:274.0RTX PRO 6000:— | MI325X:169.9RTX PRO 6000:— |
| Cost ($/M tok) | MI325X:$0.891RTX PRO 6000:— | MI325X:$1.298RTX PRO 6000:— | MI325X:$2.090RTX PRO 6000:— |
| tok/s/MW | MI325X:236377RTX PRO 6000:— | MI325X:162120RTX PRO 6000:— | MI325X:100530RTX PRO 6000:— |
| Concurrency | MI325X:~32RTX PRO 6000:— | MI325X:~16RTX PRO 6000:— | MI325X:~9RTX PRO 6000:— |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.