Provider guide
Where to run Qwen3.8 27B
18 live listings tracked — output prices vary 2.9× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
272 tok/s
measured throughput · $0.99 input / $1.49 output per million tokens
Cerebraszero-retention$0.99 in/1M$1.49 out/1M272 tok/s66Kfp162 hours agoAkashMLzero-retention$0.20 in/1M$1.78 out/1M40 tok/s262Kfp82 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.20 in/1M$0.15 in/1M$2.50 out/1Mdirect$1.88 out/1Mthrough OpenRouter26 tok/sthrough OpenRouter262Kbf162 hours agoPhalazero-retentionCHEAPEST$0.15 in/1M$1.88 out/1M86 tok/s1M—2 hours agoParasailzero-retention$0.24 in/1M$2.20 out/1M64 tok/s262Kfp82 hours agoDarkbloom$0.050 in/1M$2.20 out/1M2 tok/s262Kfp42 hours agoChutes$0.24 in/1M$2.20 out/1M30 tok/s262Kfp82 hours agoMancer 2zero-retention$0.20 in/1M$2.50 out/1M3 tok/s262Kfp82 hours agoIonstreamzero-retention$0.089 in/1M$2.50 out/1M11 tok/s262Kfp82 hours agoAlibaba Cloud$0.42 in/1M$2.55 out/1M53 tok/s1M—2 hours agoDekaLLMzero-retention$0.049 in/1M$3.00 out/1M62 tok/s262K—2 hours agoNovita AIzero-retention through OpenRouterDirect and through OpenRouter$0.42 in/1M$3.00 out/1M32 tok/sthrough OpenRouter1M—2 hours agoCoreWeavezero-retention$0.40 in/1M$3.00 out/1M51 tok/s262Kfp82 hours agoOpenRouter$0.42 in/1M$3.00 out/1M—1M—2 hours agoCloudflare Workers AI$0.45 in/1M$3.20 out/1M25 tok/s262K—14 hours agoVenice AIzero-retention$0.45 in/1M$3.20 out/1M28 tok/s262Kfp82 hours agoRekazero-retention$0.025 in/1M$4.35 out/1M6 tok/s262K—2 hours agoWaferzero-retention$0.025 in/1M$4.35 out/1M54 tok/s262K—2 hours ago
DeepInfra
Novita AI
OpenRouter
Cloudflare Workers AIFull specs, hardware verdicts and benchmarks on the Qwen3.8 27B model page →