Provider guide
Where to run DeepSeek V3.2
15 live listings tracked — output prices vary 14.5× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
GMICloud$0.21 in/1M$0.31 out/1M33 tok/s164Kfp83 days agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.26 in/1M$0.38 out/1M14 tok/sthrough OpenRouter164Kfp42 hours agoAtlasCloud$0.26 in/1M$0.38 out/1M34 tok/s164Kfp82 hours agoVenice AIzero-retention$0.27 in/1M$0.39 out/1M11 tok/s160K—2 hours agoNovita AICHEAPEST$0.27 in/1M$0.40 out/1M—164K—4 days agoOpenRouter$0.28 in/1M$0.42 out/1M—164K—2 hours agoSiliconFlowzero-retention$0.26 in/1M$0.42 out/1M14 tok/s164Kfp82 hours agoBaidu$0.28 in/1M$0.42 out/1M30 tok/s131Kfp82 hours agoDigitalOcean Gradientzero-retention$0.30 in/1M$0.96 out/1M17 tok/s164K—2 hours agoPhalazero-retention$1.00 in/1M$1.00 out/1M11 tok/s164K—2 hours agoAlibaba Cloud$0.37 in/1M$1.11 out/1M23 tok/s131Kfp82 hours agoFriendli$0.50 in/1M$1.50 out/1M45 tok/s164K—2 hours agoGoogle Vertex AIzero-retention$0.56 in/1M$1.68 out/1M9 tok/s164K—2 hours agoMarazero-retention$3.00 in/1M$4.50 out/1M21 tok/s33K—8 hours agoSambaNovazero-retention through OpenRouterDirect and through OpenRouter$3.00 in/1M$4.50 out/1M40 tok/sthrough OpenRouter33K—2 hours ago
DeepInfra
OpenRouter
Baidu
DigitalOcean Gradient
Google Vertex AI
SambaNovaFull specs, hardware verdicts and benchmarks on the DeepSeek V3.2 model page →