MiniMax M2.5/M2.7 · Chip comparison

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 13,508 tok/s/chip for $0.046 per million tokens at 31 tok/s/user on MiniMax M2.5/M2.7; MI300X posts 1,563 tok/s/chip for $0.17. B300 is 72% cheaper per token; B300 delivers 8.6× the tok/s/chip.

Throughput at 49 tok/s/user on MiniMax M2.5/M2.7: B300 hits 8,734 tok/s/chip, MI300X hits 1,396. Costs per million tokens land at $0.072 and $0.19 respectively. B300 is 62% cheaper per token; B300 delivers 6.3× the tok/s/chip.

B300 / MI300X on MiniMax M2.5/M2.7 at 67 tok/s/user: 5,346 / 1,067 tok/s/chip, $0.12 / $0.25 per million tokens. B300 is 53% cheaper per token; B300 delivers 5.0× the tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)

View performance per dollar →

Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.

Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Throughput (tok/s/chip)
B30013,508MI300X1,563
B3008,734MI300X1,396
B3005,346MI300X1,067
Cost ($/M tok)
B300$0.046MI300X$0.17
B300$0.072MI300X$0.19
B300$0.12MI300X$0.25
tok/s/MW
B3007,109,570MI300X1,124,722
B3004,596,631MI300X1,004,511
B3002,813,494MI300X767,436
Concurrency
B300~1,929MI300X~18
B300~128MI300X~6
B300~23MI300X~4