StarSeaAPI
MiniMax
M and M-highspeed families — high-throughput general models
MiniMax models are built for throughput. The highspeed variants are tuned for very low latency, which makes them a good fit for interactive products where perceived speed matters more than the last point of benchmark score.
Models (8)
| Models | Input $/M | Output $/M | Intelligence index | Coding index |
|---|---|---|---|---|
| MiniMax-M2 | $0.32 | $1.27 | — | — |
| MiniMax-M2.1 | $0.32 | $1.27 | — | — |
| MiniMax-M2.5 | $0.32 | $1.27 | — | — |
| MiniMax-M2.7 | $0.32 | $1.27 | — | 52.6 |
| MiniMax-M2.1-highspeed | $0.64 | $2.55 | — | — |
| MiniMax-M2.5-highspeed | $0.64 | $2.55 | — | — |
| MiniMax-M2.7-highspeed | $0.64 | $2.55 | — | — |
| minimax-m3 | $0.64 | $2.55 | 35.7 | 58.6 |
Prices are per 1M tokens in USD · ¥6.6 = $1