MiniMax M2.5/M2.7 — B300 vs MI355X
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) 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.
B300 / MI355X on MiniMax M2.5/M2.7 at 41 tok/s/user: 6194 / 3719 tok/s/GPU, $0.10 / $0.11 per million tokens. B300 is 5% cheaper per token; B300 delivers 67% more tok/s/GPU.
Around the middle of the 16–117 tok/s/user interactivity band, at 67 tok/s/user on MiniMax M2.5/M2.7: B300 runs 2626 tok/s/GPU at $0.25/M tokens, MI355X runs 1188 at $0.35/M. B300 is 43% cheaper per token; B300 delivers 121% more tok/s/GPU.
Setting 92 tok/s/user as the target on MiniMax M2.5/M2.7, B300 produces 1154 tok/s/GPU ($0.54 per million tokens) and MI355X produces 564 ($0.71). B300 is 31% cheaper per token; B300 delivers 105% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B300:6193.7MI355X:3718.9 | B300:2625.7MI355X:1187.6 | B300:1154.0MI355X:563.9 |
| Cost ($/M tok) | B300:$0.105MI355X:$0.110 | B300:$0.246MI355X:$0.351 | B300:$0.544MI355X:$0.714 |
| tok/s/MW | B300:2854243MI355X:1403373 | B300:1209980MI355X:448155 | B300:531800MI355X:212780 |
| Concurrency | B300:~339MI355X:~256 | B300:~80MI355X:~37 | B300:~26MI355X:~8 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.