Mistral Devstral Small 1.1 vs GPT-5.1-Codex
Compare Mistral Devstral Small 1.1 by Mistral AI against GPT-5.1-Codex by OpenAI, tested across 53 shared challenges.
Our Verdict
GPT-5.1-Codex
Mistral Devstral Small 1.1Runner-up
No community votes yet. On paper, GPT-5.1-Codex has the edge — bigger model tier, newer, bigger context window, major provider backing.
Mistral Devstral Small 1.1 is 33x cheaper per token — worth considering if cost matters.
Slight edge
API pricing
Cost per 1M tokens
Mistral Devstral Small 1.1
Input
$0.10
13× cheaper
Output
$0.30
33× cheaper
GPT-5.1-Codex
Input
$1.25
Output
$10.00
Mistral Devstral Small 1.1 is cheaper on both — 13× input, 33× output
Writing DNA
Style Comparison
Similarity
58%
GPT-5.1-Codex uses 10.4x more headings
Mistral Devstral Small 1.1
GPT-5.1-Codex
50%Vocabulary70%
17wSentence Length17w
0.56Hedging0.39
4.3Bold3.4
5.7Lists3.5
0.00Emoji0.00
0.05Headings0.50
0.34Transitions0.38
Based on 21 + 14 text responses
Research
What we learned reading every model
FAQ
