MiniMax M2.5/M2.7 — B200 vs MI325X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) 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.
B200 posts 5067 tok/s/GPU for $0.11 per million tokens at 46 tok/s/user on MiniMax M2.5/M2.7; MI325X posts 1872 tok/s/GPU for $0.19. B200 is 78% cheaper per token; B200 delivers 171% more tok/s/GPU.
Throughput at 66 tok/s/user on MiniMax M2.5/M2.7: B200 hits 2635 tok/s/GPU, MI325X hits 857. Per-million costs land at $0.21 and $0.41 respectively. B200 is 98% cheaper per token; B200 delivers 208% more tok/s/GPU.
B200 / MI325X on MiniMax M2.5/M2.7 at 85 tok/s/user: 1374 / 355 tok/s/GPU, $0.40 / $1.00 per million tokens. B200 is 153% cheaper per token; B200 delivers 287% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:5067.5MI325X:1871.6 | B200:2635.2MI325X:856.8 | B200:1373.7MI325X:354.6 |
| Cost ($/M tok) | B200:$0.107MI325X:$0.190 | B200:$0.206MI325X:$0.407 | B200:$0.396MI325X:$1.003 |
| tok/s/MW | B200:2335237MI325X:858544 | B200:1214396MI325X:393012 | B200:633031MI325X:162651 |
| Concurrency | B200:~225MI325X:~166 | B200:~82MI325X:~53 | B200:~33MI325X:~17 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.