Provider guide
Where to run DeepSeek V4 Pro
26 live listings tracked — output prices vary 2.7× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
110 tok/s
measured throughput · $1.32 input / $3.96 output per million tokens
OpenRouterCHEAPEST$0.78 in/1M$1.57 out/1M—1M—20 hours agoStreamLake$0.78 in/1M$1.57 out/1M41 tok/s1Mfp820 hours agoAlibaba Cloud$0.58 in/1M$1.74 out/1M63 tok/s1M—2 hours agoRekazero-retention$0.58 in/1M$1.74 out/1M28 tok/s1M—2 hours agoIonstreamzero-retention$1.24 in/1M$1.85 out/1M82 tok/s1M—44 hours agoGMICloud$0.96 in/1M$1.91 out/1M23 tok/s1Mfp82 hours agoDeepSeek$0.66 in/1M$1.98 out/1M9 tok/s1M—2 hours agoStreamLake$0.66 in/1M$1.98 out/1M45 tok/s1M—2 hours agoDigitalOcean Gradientzero-retention$1.04 in/1M$2.09 out/1M49 tok/s1M—2 hours agoCloudflare Workers AI$1.15 in/1M$2.55 out/1M39 tok/s1M—2 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$1.30 in/1M$2.60 out/1M33 tok/sthrough OpenRouter1Mfp82 hours agoAlibaba Cloud$1.42 in/1M$2.83 out/1M41 tok/s1Mfp88 hours agoPhalazero-retention$0.96 in/1M$2.88 out/1M49 tok/s1M—2 hours agoNovita AIzero-retention through OpenRouterDirect and through OpenRouter$1.60 in/1M$0.99 in/1M$3.20 out/1Mdirect$2.97 out/1Mthrough OpenRouter49 tok/sthrough OpenRouter1Mfp82 hours agoSiliconFlowzero-retention$1.50 in/1M$3.13 out/1M37 tok/s1Mfp82 hours agoNextBitzero-retention$1.06 in/1M$3.17 out/1M50 tok/s1Mfp82 hours agoVenice AIzero-retention$1.65 in/1M$3.30 out/1M51 tok/s1M—2 hours agoAtlasCloud$1.68 in/1M$3.38 out/1M45 tok/s1Mfp42 hours agoRelacezero-retention$0.17 in/1M$3.50 out/1M75 tok/s1Mfp42 hours agoMicrosoft Azure AIzero-retention$1.91 in/1M$3.83 out/1M58 tok/s1M—2 hours agoAtlasCloud$1.32 in/1M$3.96 out/1M34 tok/s1Mfp82 hours agoBaidu$1.32 in/1M$3.96 out/1M45 tok/s1Mfp838 hours agoParasailzero-retention$1.32 in/1M$3.96 out/1M65 tok/s1Mfp82 hours agoCoreWeavezero-retention$1.31 in/1M$3.96 out/1M60 tok/s1Mfp82 hours agoTogether AIzero-retention$1.32 in/1M$3.96 out/1M110 tok/s1M—14 hours agoWaferzero-retention$0.55 in/1M$4.20 out/1M48 tok/s1M—2 hours ago
DigitalOcean Gradient
Cloudflare Workers AI
DeepInfra
Novita AI
Microsoft Azure AI
BaiduFull specs, hardware verdicts and benchmarks on the DeepSeek V4 Pro model page →