DeepSeek V3.2vsDeepSeek V4 Pro
DeepSeek V3.2 is cheaper than DeepSeek V4 Pro at $0.28/M vs $1.74/M input tokens.
| Spec | DeepSeek V3.2 | DeepSeek V4 Pro |
|---|---|---|
| Input price | $0.28/M tokens | $1.74/M tokens |
| Output price | $0.42/M tokens | $3.48/M tokens |
| Context window | 131K tokens | 1.0M tokens |
| Weights | Open | Open |
| Free API (OpenRouter) | No | No |
| Released | Dec 2025 | Apr 2026 |
Our Verdict
DeepSeek V4 Pro
DeepSeek V3.2Runner-up
Not enough votes to call it. On the specs, DeepSeek V4 Pro has the edge: bigger model tier, newer, bigger context window.
DeepSeek V3.2 costs 8.3x less per token.
Slight edge
Reviewing agent-written code?See a Brief PR report
API pricing
Cost per 1M tokens
DeepSeek V3.2
Input
$0.28
6.2× cheaper
Output
$0.42
8.3× cheaper
DeepSeek V4 Pro
Input
$1.74
Output
$3.48
DeepSeek V3.2 is cheaper on both: 6.2× input, 8.3× output.
Where to run it
30 hosts, cheapest first
DeepSeek V3.215 hosts
GMI Cloudfp8$0.21 in$0.31 out164k99.9% upStreamLakefp8$0.21 in$0.32 out128k99.6% upDigitalOcean$0.25 in$0.80 out164k97.8% upSiliconFlowfp8$0.26 in$0.42 out164k98.2% upAtlasCloudfp8$0.26 in$0.38 out164k99.8% upDeepInfrafp4$0.26 in$0.38 out164k99.5% up
9 more hostsFewer hosts
Novitafp8$0.27 in$0.40 out164k100% upBaidu Qianfanfp8$0.28 in$0.42 out131k100% upVenice$0.33 in$0.48 out160k98.7% upAlibaba Cloudfp8$0.37 in$1.11 out131k99.4% upFriendli$0.50 in$1.50 out164k100% upGoogle Vertex AI$0.56 in$1.68 out164k99.8% upPhala$1.00 in$1.00 out164k99.1% upMara$3.00 in$4.50 out33k75% upSambaNova$3.00 in$4.50 out33k89% up
DeepSeek V4 Pro15 hosts
DigitalOcean$0.87 in$1.74 out1M97% upBaidu Qianfanfp8$0.96 in$1.91 out1M99.7% upGMI Cloudfp8$0.96 in$1.91 out1M99.6% upDeepInfrafp8$1.30 in$2.60 out1M99.2% upAlibaba Cloudfp8$1.42 in$2.83 out1M100% upSiliconFlowfp8$1.50 in$3.13 out1M97.9% up
9 more hostsFewer hosts
Novitafp8$1.60 in$3.20 out1M100% upAtlasCloudfp4$1.68 in$3.38 out1M99.1% upBasetenfp4$1.74 in$3.48 out1M98.1% upParasailfp8$1.74 in$3.48 out1M99% upAzure AI Foundry$1.91 in$3.83 out1M98.3% upNextBitfp8$2.00 in$4.00 out1M99.8% upStreamLakefp8degraded$0.95 in$1.90 out1M96.7% upFireworksdegraded$1.20 in$1.20 out1M0% upVenicedegraded$1.65 in$3.30 out1M96.2% up
Per million tokens. Prices and uptime via OpenRouter, checked 11 Sep 2026.
Writing DNA
Style Comparison
Similarity
38%
DeepSeek V4 Pro uses 3.5x more emoji
DeepSeek V3.2
DeepSeek V4 Pro
54%Vocabulary57%
29wSentence Length18w
0.58Hedging0.52
5.9Bold3.3
4.8Lists2.3
0.00Emoji0.04
0.74Headings0.45
0.12Transitions0.15
Based on 28 + 27 text responses
Research
What we learned reading every model
FAQ
Common questions
Keep exploring
More comparisons
DeepSeek V3.2 vs Ox AlphaNew provider
DeepSeek V3.2 vs MiniMax M3New provider
DeepSeek V3.2 vs MiniMax M2.7New provider
DeepSeek V3.2 vs Aurora AlphaNew provider
DeepSeek V3.2 vs Claude 2New provider
DeepSeek V3.2 vs Claude 3 HaikuNew provider
DeepSeek V3.2 vs Claude 3 OpusNew provider
DeepSeek V3.2 vs Claude 3 SonnetNew provider