Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and H100 (NVIDIA Hopper) on MiniMax M2.5/M2.7. 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.
At 44 tok/s/user interactivity on MiniMax M2.5/M2.7, B200 delivers 9526 tok/s/chip at $0.05 per million tokens; H100 delivers 1891 tok/s/chip at $0.17. B200 is 241% cheaper per token; B200 delivers 404% more tok/s/chip at this point.
B200 posts 4753 tok/s/chip for $0.10 per million tokens at 66 tok/s/user on MiniMax M2.5/M2.7; H100 posts 1269 tok/s/chip for $0.26. B200 is 153% cheaper per token; B200 delivers 275% more tok/s/chip.
Throughput at 89 tok/s/user on MiniMax M2.5/M2.7: B200 hits 3180 tok/s/chip, H100 hits 806. Per-million costs land at $0.15 and $0.40 respectively. B200 is 167% cheaper per token; B200 delivers 295% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:9525.9H100:1891.3 | B200:4752.7H100:1268.8 | B200:3179.8H100:805.9 |
| Cost ($/M tok) | B200:$0.050H100:$0.172 | B200:$0.101H100:$0.256 | B200:$0.151H100:$0.403 |
| tok/s/MW | B200:5570725H100:1380510 | B200:2779362H100:926145 | B200:1859520H100:588262 |
| Concurrency | B200:~512H100:~39 | B200:~21H100:~18 | B200:~20H100:~9 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr: VR200: 3.61H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68Jalapeño (Teacup): 1.47
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.
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
MiniMax M2.5/2.7 230B 8K / 1K Total Tokens per $1 TCO vs. Interactivity