Mistral Devstral Small 1.1 vs Kimi K2 Thinking
Compare Mistral Devstral Small 1.1 by Mistral AI against Kimi K2 Thinking by Moonshot AI, tested across 54 shared challenges.
Our Verdict
Kimi K2 Thinking
Mistral Devstral Small 1.1Runner-up
No community votes yet. On paper, Kimi K2 Thinking has the edge — bigger model tier, newer, bigger context window.
Mistral Devstral Small 1.1 is 8.3x cheaper per token — worth considering if cost matters.
Too close to call
API pricing
Cost per 1M tokens
Mistral Devstral Small 1.1
Input
$0.10
6.0× cheaper
Output
$0.30
8.3× cheaper
Kimi K2 Thinking
Input
$0.60
Output
$2.50
Mistral Devstral Small 1.1 is cheaper on both — 6.0× input, 8.3× output
Writing DNA
Style Comparison
Similarity
55%
Kimi K2 Thinking uses 11.7x more headings
Mistral Devstral Small 1.1
Kimi K2 Thinking
50%Vocabulary57%
17wSentence Length26w
0.56Hedging0.36
4.3Bold5.8
5.7Lists6.0
0.00Emoji0.00
0.05Headings0.56
0.34Transitions0.06
Based on 21 + 14 text responses
Research
What we learned reading every model
FAQ
