MiniMax M2.5/M2.7 — GB300 NVL72 vs H100
Head-to-head AI inference benchmark comparison of GB300 NVL72 (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.
GB300 NVL72 posts 3985 tok/s/GPU for $0.19 per million tokens at 59 tok/s/user on MiniMax M2.5/M2.7; H100 posts 616 tok/s/GPU for $0.58. GB300 NVL72 is 207% cheaper per token; GB300 NVL72 delivers 547% more tok/s/GPU.
Throughput at 78 tok/s/user on MiniMax M2.5/M2.7: GB300 NVL72 hits 1610 tok/s/GPU, H100 hits 380. Per-million costs land at $0.48 and $0.94 respectively. GB300 NVL72 is 96% cheaper per token; GB300 NVL72 delivers 324% more tok/s/GPU.
GB300 NVL72 / H100 on MiniMax M2.5/M2.7 at 98 tok/s/user: 901 / 210 tok/s/GPU, $0.84 / $1.72 per million tokens. GB300 NVL72 is 103% cheaper per token; GB300 NVL72 delivers 328% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | GB300 NVL72:3984.8H100:616.0 | GB300 NVL72:1610.0H100:379.6 | GB300 NVL72:901.3H100:210.4 |
| Cost ($/M tok) | GB300 NVL72:$0.187H100:$0.576 | GB300 NVL72:$0.479H100:$0.936 | GB300 NVL72:$0.845H100:$1.717 |
| tok/s/MW | GB300 NVL72:1897520H100:356072 | GB300 NVL72:766668H100:219420 | GB300 NVL72:429188H100:121604 |
| Concurrency | GB300 NVL72:~405H100:~43 | GB300 NVL72:~64H100:~12 | GB300 NVL72:~32H100:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.