Short answer: Grok 4.7 is 1.7x cheaper (0.79¢ vs 1.36¢ a chat). Kimi K3 ranks higher overall (#8 vs #9). Kimi K3 wins at coding, maths. Grok 4.7 wins at reasoning. Pick Grok 4.7 for everyday use and Kimi K3 when quality matters most.
MMOONSHOTGANG
KIMI K3
MOONSHOT GANG
CRAPGPT.LOLINDEPENDENT AI MODEL REVIEWS
CRAP SCORE563 SOURCES
REACHFURTHERTHINKLONGER
LONGCONTEXTDEEPORBIT
POWER74EPOCH · AA · CRAPGPT
Reasoning
76
Code
76
Math
73
Agentic
69
Instruction
54
Value
22
BOSS★★★★
REASONER / CODER
1.36¢ per chat · 73 chats per NZ$1
“Deep thought. Low drama.”
Every model is somebody.
— CRAPGPT.LOL —
XXAIGANG
GROK 4.7
XAI GANG
CRAPGPT.LOLINDEPENDENT AI MODEL REVIEWS
CRAP SCORE533 SOURCES
REALQUESTIONS.BOLDERANSWERS.
HIGHERCEILINGS.HARDERPROBLEMS.
POWER68EPOCH · AA · CRAPGPT
Reasoning
79
Code
69
Writing
54
Math
44
Instruction
4
Value
26
BOSS★★★★
REASONER / CODER
0.79¢ per chat · 126 chats per NZ$1
“See the whole pattern.”
Every model is somebody.
— CRAPGPT.LOL —
Kimi K3
Grok 4.7
Cost per chat (NZD)
1.36¢
0.79¢
Chats per NZ$1
73
126
Overall rank
#8
#9
Vendor
Moonshot AI
xAI
Speciality
Reasoning
Reasoning
Reasoning
76
79
Code
76
69
Math
73
44
Scores are 0 to 100 percentiles blending LMArena, Epoch AI, Artificial Analysis and CrapGPT's own tests. Updated 2026-10-13.
FAQ
Which is cheaper, Kimi K3 or Grok 4.7?
Grok 4.7 costs 0.79¢ per typical chat versus 1.36¢ for Kimi K3 on CrapGPT (NZD, measured token use).
Which is better, Kimi K3 or Grok 4.7?
Kimi K3 ranks higher overall on CrapGPT's blended benchmark score. By category: Kimi K3 leads at coding, maths; Grok 4.7 leads at reasoning.
Can I try Kimi K3 and Grok 4.7 side by side?
Yes. CrapGPT's card battle sends one prompt to two models and lets you pick the winner, paying only for the two replies.
