BENCHMARKFirst benchmarks recorded: MiniMax M3 → Cortecs → ?6hBENCHMARKFirst benchmarks recorded: Kimi K3 → Cortecs → ?6hBENCHMARKFirst benchmarks recorded: DeepSeek V4 Pro → Cortecs → ?6hBENCHMARKFirst benchmarks recorded: GLM-4.7 → Cortecs → ?6hBENCHMARKFirst benchmarks recorded: GLM-5.2 → Cortecs → ?6h

DeepSeek-V3

6 providers · 671B MoE · 128k context · 3 regions · all models →

DeepSeek-V3 provider ranking

Updated 12m agoDEV DATA
#ProviderΔ30d
01NebiusBEST OVERALLEU RESIDENTNL · Serverless · DeepSeek-V398.9%97.7%99.1%34012573.8s116$0.16$0.50↑ 1.9
02FireworksMOST RELIABLEUS · Serverless · DeepSeek-V399.0%98.9%99.7%36812723.0s153$0.28$1.10↑ 1.0
03TogetherUS · Serverless · DeepSeek-V397.8%96.8%98.3%47316004.2s107$0.30$0.90↑ 0.1
04OpenRouterUS · Routed · DeepSeek-V397.6%98.6%98.7%45314653.5s131$0.28$1.10↑ 1.2
05DeepInfraUS · Serverless · DeepSeek-V397.4%96.1%98.0%60416966.2s72$0.10$0.45↓ 1.3
06SiliconFlowBEST VALUECN · Serverless · DeepSeek-V396.6%94.8%97.7%69320486.0s75$0.14$0.39↑ 0.7
6 of 6 providers ranked · Methodology →
DeepSeek-V3 inference benchmark: fastest & cheapest providers · InferenceBench