Short answer: Gemini 3.8 Flash is 7.8x cheaper (0.43¢ vs 3.35¢ a chat). Claude Opus 5 ranks higher overall (#2 vs #4). Claude Opus 5 wins at coding, maths, writing, following instructions, agentic tasks. Gemini 3.8 Flash wins at reasoning. Pick Gemini 3.8 Flash for everyday use and Claude Opus 5 when quality matters most.
AANTHROPICGANG
CLAUDE OPUS 5
ANTHROPIC GANG
CRAPGPT.LOLINDEPENDENT AI MODEL REVIEWS
CRAP SCORE604 SOURCES
THINKDEEPERWRITECLEANERBUILD SAFER
SAFERIDEASA MOREHUMANTOMORROW
POWER89ARENA · EPOCH · AA · CRAPGPT
Writing
93
Agentic
92
Reasoning
90
Code
88
Instruction
88
Value
6
LIEUTENANT★★★★★
GENERALIST / WRITER
3.35¢ per chat · 29 chats per NZ$1
“Ask me anything. Pay less.”
Every model is somebody.
— CRAPGPT.LOL —
GGOOGLEGANG
GEMINI 3.8 FLASH
GOOGLE GANG
CRAPGPT.LOLINDEPENDENT AI MODEL REVIEWS
CRAP SCORE684 SOURCES
MULTIMODALINTELLIGENCEFOR A MOREOPENTOMORROW
SEEUNDERSTANDREASONCREATE
POWER82ARENA · EPOCH · AA · CRAPGPT
Reasoning
91
Writing
88
Instruction
85
Agentic
85
Code
78
Value
42
OPERATOR★★★★
GENERALIST / REASONER
0.43¢ per chat · 232 chats per NZ$1
“Polished. Dangerous. Usually right.”
Every model is somebody.
— CRAPGPT.LOL —
Claude Opus 5
Gemini 3.8 Flash
Cost per chat (NZD)
3.35¢
0.43¢
Chats per NZ$1
29
232
Overall rank
#2
#4
Vendor
Anthropic
Google
Speciality
General
General
Reasoning
90
91
Code
88
78
Math
86
69
Writing
93
88
Instruction
88
85
Agentic
92
85
Scores are 0 to 100 percentiles blending LMArena, Epoch AI, Artificial Analysis and CrapGPT's own tests. Updated 2026-10-11.
FAQ
Which is cheaper, Claude Opus 5 or Gemini 3.8 Flash?
Gemini 3.8 Flash costs 0.43¢ per typical chat versus 3.35¢ for Claude Opus 5 on CrapGPT (NZD, measured token use).
Which is better, Claude Opus 5 or Gemini 3.8 Flash?
Claude Opus 5 ranks higher overall on CrapGPT's blended benchmark score. By category: Claude Opus 5 leads at coding, maths, writing, following instructions, agentic tasks; Gemini 3.8 Flash leads at reasoning.
Can I try Claude Opus 5 and Gemini 3.8 Flash 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.
