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 4567 tok/s/chip for $0.11 per million tokens at 46 tok/s/user on MiniMax M2.5/M2.7; MI325X posts 1643 tok/s/chip for $0.19. B200 is 73% cheaper per token; B200 delivers 178% more tok/s/chip.
Throughput at 65 tok/s/user on MiniMax M2.5/M2.7: B200 hits 2346 tok/s/chip, MI325X hits 757. Per-million costs land at $0.21 and $0.41 respectively. B200 is 101% cheaper per token; B200 delivers 210% more tok/s/chip.
B200 / MI325X on MiniMax M2.5/M2.7 at 85 tok/s/user: 1382 / 325 tok/s/chip, $0.35 / $0.94 per million tokens. B200 is 169% cheaper per token; B200 delivers 326% more tok/s/chip. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:4566.6MI325X:1642.6 | B200:2346.1MI325X:757.2 | B200:1381.7MI325X:324.7 |
| Cost ($/M tok) | B200:$0.108MI325X:$0.186 | B200:$0.205MI325X:$0.412 | B200:$0.350MI325X:$0.940 |
| tok/s/MW | B200:2670531MI325X:971943 | B200:1371990MI325X:448067 | B200:808011MI325X:192115 |
| Concurrency | B200:~388MI325X:~147 | B200:~109MI325X:~38 | B200:~33MI325X:~16 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.