MiniMax M3 428B — H200 vs Vera Rubin NVL72 Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus Vera Rubin NVL72 (NVIDIA Vera Rubin) on MiniMax M3 428B. 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 77 tok/s/user on MiniMax M3 428B, H200 comes in at $0.49 per million tokens. Vera Rubin NVL72 hasn't been benchmarked at this operating point.
Only H200 has cost data at 149 tok/s/user on MiniMax M3 428B — $1.00 per million tokens. Vera Rubin NVL72 is unmeasured at this target.
H200 costs $2.01 per million tokens at 222 tok/s/user on MiniMax M3 428B; we have no Vera Rubin NVL72 benchmark data at this exact target. (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.)
Chip pricing (owning hyperscaler): H200 $1.22/chip/hr · Vera Rubin NVL72 $3.61/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | H200:$0.488Vera Rubin NVL72:— | H200:$1.002Vera Rubin NVL72:— | H200:$2.010Vera Rubin NVL72:— |
| Concurrency | H200:~20Vera Rubin NVL72:— | H200:~4Vera Rubin NVL72:— | H200:~3Vera Rubin NVL72:— |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
MiniMax M3 428B 8K / 1K Cost per Million Total Tokens vs. Interactivity
MiniMax M3 428B 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
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip