InferenceX
中文
Qwen 3.5 397B-A17B · Performance per Dollar

Qwen 3.5 397B-A17B — MI300X vs RTX PRO 6000 Performance per Dollar

Cost per million tokens of MI300X (AMD CDNA 3) versus RTX PRO 6000 (NVIDIA Blackwell) on Qwen 3.5 397B-A17B. Large-hyperscaler-volume ownership TCO normalized by total 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.

At 47 tok/s/user on Qwen 3.5 397B-A17B, RTX PRO 6000 comes in at $0.18 per million tokens. MI300X hasn't been benchmarked at this operating point.

Only RTX PRO 6000 has cost data at 82 tok/s/user on Qwen 3.5 397B-A17B — $0.24 per million tokens. MI300X is unmeasured at this target.

RTX PRO 6000 costs $0.35 per million tokens at 118 tok/s/user on Qwen 3.5 397B-A17B; we have no MI300X benchmark data at this exact target. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)

Chip pricing (owning hyperscaler): MI300X $0.95/chip/hr · RTX PRO 6000 $0.68/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.

View full latency + throughput comparison →

Qwen 3.5 397B-A17B: MI300X versus RTX PRO 6000 cost per million tokens at matched interactivity levels
MI300X versus RTX PRO 6000 cost per million tokens for this comparison's canonical default workload. Lower cost indicates better performance per dollar.
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Dollar per Million Tokens
MI300X:—RTX PRO 6000:$0.177
MI300X:—RTX PRO 6000:$0.237
MI300X:—RTX PRO 6000:$0.351
Concurrency
MI300X:—RTX PRO 6000:~12
MI300X:—RTX PRO 6000:~5
MI300X:—RTX PRO 6000:~2

Inference Performance

Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.

Benchmark Config

Qwen3.5 397B 8K / 1K Cost per Million Total Tokens vs. Interactivity

Cost Tier:
Source:
SemiAnalysis InferenceX™

TCO $/chip/hr:

Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model

Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.

Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.

Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.

Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.

Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.

No data available

Matching measurements exist, but their chip series are hidden.

Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip