Short answer: MiMo V2.5 is 44x cheaper (0.018¢ vs 0.79¢ a chat). Grok 4.7 ranks higher overall (#15 vs #17). Grok 4.7 wins at reasoning, coding, writing. MiMo V2.5 wins at maths, following instructions. Pick MiMo V2.5 for everyday use and Grok 4.7 when quality matters most.


| Grok 4.7 | MiMo V2.5 | |
|---|---|---|
| Cost per chat (NZD) | 0.79¢ | 0.018¢ |
| Chats per NZ$1 | 126 | 5,549 |
| Overall rank | #15 | #17 |
| Vendor | xAI | Xiaomi |
| Speciality | Reasoning | General |
| Crowd Elo | no votes | 1000 |
| Reasoning | 71 | 66 |
| Code | 64 | 60 |
| Math | 47 | 53 |
| Writing | 61 | 52 |
| Instruction | 47 | 64 |
Scores are 0 to 100 percentiles blending LMArena, Epoch AI, Artificial Analysis and CrapGPT's own tests. Updated 2026-10-11.
MiMo V2.5 costs 0.018¢ per typical chat versus 0.79¢ for Grok 4.7 on CrapGPT (NZD, measured token use).
Grok 4.7 ranks higher overall on CrapGPT's blended benchmark score. By category: Grok 4.7 leads at reasoning, coding, writing; MiMo V2.5 leads at maths, following instructions.
Yes. CrapGPT's card battle sends one prompt to two models and lets you pick the winner, paying only for the two replies.