MiniMax M2.5/M2.7 · GPU comparison

MiniMax M2.5/M2.7 — H100 vs MI355X

Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI355X (AMD CDNA 4) 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.

H100 posts 628 tok/s/GPU for $0.57 per million tokens at 59 tok/s/user on MiniMax M2.5/M2.7; MI355X posts 1613 tok/s/GPU for $0.27. MI355X is 115% cheaper per token; MI355X delivers 157% more tok/s/GPU.

Throughput at 78 tok/s/user on MiniMax M2.5/M2.7: H100 hits 376 tok/s/GPU, MI355X hits 828. Per-million costs land at $0.94 and $0.48 respectively. MI355X is 94% cheaper per token; MI355X delivers 121% more tok/s/GPU.

H100 / MI355X on MiniMax M2.5/M2.7 at 97 tok/s/user: 208 / 490 tok/s/GPU, $1.75 / $0.82 per million tokens. MI355X is 115% cheaper per token; MI355X delivers 136% 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.)

View performance-per-dollar view →

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)
Throughput (tok/s/gpu)
H100:628.0MI355X:1612.8
H100:375.6MI355X:828.3
H100:207.9MI355X:489.7
Cost ($/M tok)
H100:$0.572MI355X:$0.266
H100:$0.939MI355X:$0.485
H100:$1.754MI355X:$0.815
tok/s/MW
H100:363025MI355X:608616
H100:217088MI355X:312550
H100:120199MI355X:184802
Concurrency
H100:~43MI355X:~56
H100:~12MI355X:~15
H100:~8MI355X:~8

Inference Performance

Inference performance metrics across different models, hardware configurations, and serving parameters.