Head to head
MiniMax M2.7 vs GPT-5.4 nano
On a blended workload of one million input tokens and three million output tokens, MiniMax M2.7 is line-ball with GPT-5.4 nano.
| MiniMax M2.7 | GPT-5.4 nano | |
|---|---|---|
| Provider | Fireworks AI | OpenAI |
| Input / MTok | $0.3 | $0.2 |
| Output / MTok | $1.2 | $1.25 |
| Cache read / MTok | $0.06 | $0.02 |
| Context window | — | — |
| Blended 1M in + 3M out | $3.90 | $3.95 |
| Status | ga | ga |