Qwen 3.5 397B-A17B — B300 vs MI355X Performance per Dollar
Cost per million tokens of B300 (NVIDIA Blackwell) versus MI355X (AMD CDNA 4) on Qwen 3.5 397B-A17B. Owning-hyperscaler TCO normalized by output tokens — performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
Push Qwen 3.5 397B-A17B to 81 tok/s/user and B300 lands at $0.27 per million tokens against MI355X's $0.26 — MI355X pulls ahead by 6%.
B300: $0.56 per million tokens. MI355X: $0.45. Both at 133 tok/s/user on Qwen 3.5 397B-A17B, with MI355X 25% cheaper.
Toward the upper edge of the 30–236 tok/s/user interactivity band — at 185 tok/s/user — B300 runs $0.86 per million tokens on Qwen 3.5 397B-A17B while MI355X runs $0.78. MI355X is the cheaper choice by 11%. (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.)
GPU pricing (owning hyperscaler): B300 $2.34/GPU/hr · MI355X $1.48/GPU/hr. Source: SemiAnalysis Market August 2025 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B300:$0.272MI355X:$0.257 | B300:$0.564MI355X:$0.453 | B300:$0.865MI355X:$0.781 |
| Concurrency | B300:~61MI355X:~41 | B300:~18MI355X:~14 | B300:~9MI355X:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.