MiniMax M3 428B — B200 vs MI300X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI300X (AMD CDNA 3) on MiniMax M3 428B. 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.
Setting 47 tok/s/user as the target on MiniMax M3 428B, B200 produces 3397 tok/s/GPU ($0.16 per million tokens) and MI300X produces 517 ($0.59). B200 is 263% cheaper per token; B200 delivers 557% more tok/s/GPU.
At 81 tok/s/user interactivity on MiniMax M3 428B, B200 delivers 1700 tok/s/GPU at $0.32 per million tokens; MI300X delivers 202 tok/s/GPU at $1.53. B200 is 380% cheaper per token; B200 delivers 741% more tok/s/GPU at this point.
B200 posts 1396 tok/s/GPU for $0.41 per million tokens at 115 tok/s/user on MiniMax M3 428B; MI300X posts 142 tok/s/GPU for $2.14. B200 is 424% cheaper per token; B200 delivers 880% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:3396.9MI300X:517.4 | B200:1700.4MI300X:202.2 | B200:1395.8MI300X:142.4 |
| Cost ($/M tok) | B200:$0.162MI300X:$0.590 | B200:$0.319MI300X:$1.534 | B200:$0.409MI300X:$2.145 |
| tok/s/MW | B200:1986498MI300X:372210 | B200:994389MI300X:145436 | B200:816286MI300X:102431 |
| Concurrency | B200:~175MI300X:~47 | B200:~63MI300X:~11 | B200:~28MI300X:~5 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.