MiniMax M2.5/M2.7 — B300 vs MI300X
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and MI300X (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.
B300 posts 5614 tok/s/GPU for $0.12 per million tokens at 44 tok/s/user on MiniMax M2.5/M2.7; MI300X posts 1249 tok/s/GPU for $0.25. B300 is 116% cheaper per token; B300 delivers 350% more tok/s/GPU.
Throughput at 61 tok/s/user on MiniMax M2.5/M2.7: B300 hits 3178 tok/s/GPU, MI300X hits 779. Per-million costs land at $0.20 and $0.40 respectively. B300 is 97% cheaper per token; B300 delivers 308% more tok/s/GPU.
B300 / MI300X on MiniMax M2.5/M2.7 at 78 tok/s/user: 1799 / 412 tok/s/GPU, $0.37 / $0.75 per million tokens. B300 is 103% cheaper per token; B300 delivers 337% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B300:5614.4MI300X:1248.6 | B300:3178.5MI300X:778.8 | B300:1799.0MI300X:411.9 |
| Cost ($/M tok) | B300:$0.116MI300X:$0.250 | B300:$0.204MI300X:$0.402 | B300:$0.367MI300X:$0.747 |
| tok/s/MW | B300:2587270MI300X:697553 | B300:1464746MI300X:435091 | B300:829013MI300X:230105 |
| Concurrency | B300:~265MI300X:~63 | B300:~106MI300X:~26 | B300:~47MI300X:~11 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.