Head to head
GPT-5.4 nano vs MiniMax M2.7
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.
| GPT-5.4 nano | MiniMax M2.7 | |
|---|---|---|
| Provider | OpenAI | Fireworks AI |
| Input / MTok | $0.2 | $0.3 |
| Output / MTok | $1.25 | $1.2 |
| Cache read / MTok | $0.02 | $0.06 |
| Context window | — | — |
| Blended 1M in + 3M out | $3.95 | $3.90 |
| Status | ga | ga |