Claude Opus 5 vs Gemini 2.5 Flash Preview (thinking)
Compare Claude Opus 5 by Anthropic against Gemini 2.5 Flash Preview (thinking) by Google AI, context windows of 1.0M vs 1.0M, tested across 20 shared challenges.
Our Verdict
Claude Opus 5
Gemini 2.5 Flash Preview (thinking)Runner-up
No community votes yet. On paper, Claude Opus 5 has the edge — bigger model tier, newer.
Gemini 2.5 Flash Preview (thinking) is 7.1x cheaper per token — worth considering if cost matters.
Too close to call
API pricing
Cost per 1M tokens
Claude Opus 5
Input
$5.00
Output
$25.00
Gemini 2.5 Flash Preview (thinking)
Input
$0.17
29× cheaper
Output
$3.50
7.1× cheaper
Gemini 2.5 Flash Preview (thinking) is cheaper on both — 29× input, 7.1× output
Writing DNA
Style Comparison
Similarity
52%
Claude Opus 5 uses 1.7x more emoji
Claude Opus 5
Gemini 2.5 Flash Preview (thinking)
54%Vocabulary53%
21wSentence Length14w
0.40Hedging0.38
3.9Bold4.4
1.9Lists4.1
0.02Emoji0.00
0.71Headings0.00
0.05Transitions0.24
Based on 27 + 7 text responses
Research
What we learned reading every model
FAQ
