MiniMax M2.5/M2.7 · Chip comparison

MiniMax M2.5/M2.7 — B300 vs MI325X

Head-to-head AI inference benchmark comparison of B300 (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.

At 35 tok/s/user interactivity on MiniMax M2.5/M2.7, B300 delivers 11,471 tok/s/chip at $0.055 per million tokens; MI325X delivers 2,647 tok/s/chip at $0.12. B300 is 53% cheaper per token; B300 delivers 4.3× the tok/s/chip at this point.

B300 posts 7,774 tok/s/chip for $0.081 per million tokens at 56 tok/s/user on MiniMax M2.5/M2.7; MI325X posts 1,814 tok/s/chip for $0.17. B300 is 52% cheaper per token; B300 delivers 4.3× the tok/s/chip.

Throughput at 78 tok/s/user on MiniMax M2.5/M2.7: B300 hits 4,363 tok/s/chip, MI325X hits 1,009. Costs per million tokens land at $0.14 and $0.30 respectively. B300 is 52% cheaper per token; B300 delivers 4.3× 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)
B30011,471MI325X2,647
B3007,774MI325X1,814
B3004,363MI325X1,009
Cost ($/M tok)
B300$0.055MI325X$0.12
B300$0.081MI325X$0.17
B300$0.14MI325X$0.30
tok/s/MW
B3006,037,520MI325X1,566,421
B3004,091,823MI325X1,073,566
B3002,296,114MI325X597,118
Concurrency
B300~521MI325X~18
B300~335MI325X~8
B300~6MI325X~5