DeepSeek V3.1vsKimi K2
DeepSeek V3.1 is cheaper than Kimi K2 at $0.2/M vs $0.57/M input tokens.
| Spec | DeepSeek V3.1 | Kimi K2 |
|---|---|---|
| Input price | $0.2/M tokens | $0.57/M tokens |
| Output price | $0.8/M tokens | $2.3/M tokens |
| Context window | 164K tokens | — |
| Released | Aug 2025 | Jul 2025 |
| SWE-bench Verified | 66.0% | 65.8% |
Our Verdict
DeepSeek V3.1
Kimi K2Runner-up
Not enough votes to call it. On the specs, DeepSeek V3.1 has the edge: newer, bigger context window, major provider backing.
Kimi K2 takes Reasoning.
Too close to call
API pricing
Cost per 1M tokens
DeepSeek V3.1
Input
$0.20
2.8× cheaper
Output
$0.80
2.9× cheaper
Kimi K2
Input
$0.57
Output
$2.30
DeepSeek V3.1 is cheaper on both — 2.8× input, 2.9× output
Writing DNA
Style Comparison
Similarity
33%
Kimi K2 uses 29.1x more emoji
DeepSeek V3.1
Kimi K2
53%Vocabulary63%
14wSentence Length19w
0.42Hedging0.35
4.0Bold4.2
3.6Lists3.6
0.02Emoji0.68
0.40Headings0.61
0.22Transitions0.06
Based on 23 + 22 text responses
Research
What we learned reading every model
FAQ
