MiniMax: MiniMax M3 vs OpenAI: GPT-5.2-Codex
MiniMax: MiniMax M3 wins on more metrics (9 of 9), but the right pick depends on what you optimize for — see the breakdown below.
| Metric | MiniMax: MiniMax M3 | OpenAI: GPT-5.2-Codex |
|---|---|---|
| Intelligence Index | 44.4✓ | 40.1 |
| Coding Index | 58.6✓ | — |
| GPQA Diamond | 93%✓ | 90% |
| Design Arena Elo | 1305✓ | — |
| Speed (tokens/sec) | 73✓ | 64 |
| Latency | 828ms✓ | 1.6s |
| Input price /M | $0.300✓ | $1.75 |
| Output price /M | $1.20✓ | $14.00 |
| Context window | 1.0M✓ | 400K |
| Capabilities | ReasoningToolsJSONVision | ReasoningToolsJSONVision |