MiniMax M2.5/M2.7 — H100 vs H200
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and H200 (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 59 tok/s/user interactivity on MiniMax M2.5/M2.7, H100 delivers 628 tok/s/GPU at $0.57 per million tokens; H200 delivers 948 tok/s/GPU at $0.41. H200 is 39% cheaper per token; H200 delivers 51% more tok/s/GPU at this point.
H100 posts 376 tok/s/GPU for $0.94 per million tokens at 78 tok/s/user on MiniMax M2.5/M2.7; H200 posts 438 tok/s/GPU for $0.86. H200 is 9% cheaper per token; H200 delivers 17% more tok/s/GPU.
Throughput at 97 tok/s/user on MiniMax M2.5/M2.7: H100 hits 208 tok/s/GPU, H200 hits 277. Per-million costs land at $1.75 and $1.41 respectively. H200 is 24% cheaper per token; H200 delivers 33% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | H100:628.0H200:947.9 | H100:375.6H200:438.3 | H100:207.9H200:277.3 |
| Cost ($/M tok) | H100:$0.572H200:$0.412 | H100:$0.939H200:$0.863 | H100:$1.754H200:$1.413 |
| tok/s/MW | H100:363025H200:547943 | H100:217088H200:253368 | H100:120199H200:160316 |
| Concurrency | H100:~43H200:~66 | H100:~12H200:~23 | H100:~8H200:~12 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.