Head to head
MiniMax M3 vs GPT-5.6 Luna
On a blended workload of one million input tokens and three million output tokens, GPT-5.6 Luna is 1.0× cheaper than MiniMax M3.
| MiniMax M3 | GPT-5.6 Luna | |
|---|---|---|
| Provider | Together AI | OpenAI |
| Input / MTok | $0.3 | $0.2 |
| Output / MTok | $1.2 | $1.2 |
| Cache read / MTok | — | $0.02 |
| Context window | 524,288 | — |
| Blended 1M in + 3M out | $3.90 | $3.80 |
| Status | ga | ga |