LLMap
Models /DeepSeek V3.1 /Where to run
Provider guide

Where to run DeepSeek V3.1

8 live listings tracked — output prices vary 1.8× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.

CHEAPEST
OpenRouter
$0.25 / $0.95
per 1M tokens in / out
FASTEST MEASURED
Mara
94 tok/s
measured throughput · $0.60 input / $1.70 output per million tokens
DeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.25 in/1M$0.95 out/1M6 tok/sthrough OpenRouter164Kfp43 hours agoOpenRouterCHEAPEST$0.25 in/1M$0.95 out/1M—164K—3 hours agoSiliconFlowzero-retention$0.27 in/1M$1.00 out/1M17 tok/s164Kfp83 hours agoNovita AI$0.27 in/1M$1.00 out/1M—131K—4 days agoAtlasCloud$0.30 in/1M$1.00 out/1M44 tok/s131Kfp83 hours agoSambaNovazero-retention through OpenRouterDirect and through OpenRouter$3.00 in/1M$0.65 in/1M$4.50 out/1Mdirect$1.50 out/1Mthrough OpenRouter45 tok/sthrough OpenRouter131Kfp83 hours agoCoreWeavezero-retention$0.55 in/1M$1.65 out/1M50 tok/s161Kfp83 hours agoMarazero-retention$0.60 in/1M$1.70 out/1M94 tok/s131K—3 hours ago
Full specs, hardware verdicts and benchmarks on the DeepSeek V3.1 model page →