Provider guide
Where to run Qwen3.6 27B
13 live offers tracked — output prices vary 1.8× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
92 tok/s
measured throughput · $0.60 input / $3.60 output per million tokens
Chutes$0.30 in/1M$2.00 out/1M30 tok/s262Kfp815m agoOpenRouterCHEAPEST$0.29 in/1M$2.40 out/1M—262K—17m agoAlibaba Cloud$0.45 in/1M$2.70 out/1M57 tok/s262Kfp815m agoAlibaba Cloud$0.45 in/1M$2.70 out/1M33 tok/s262K—6d agoDeepInfra$0.32 in/1M$3.20 out/1M—262Kfp86h agoNovita AI$0.60 in/1M$3.60 out/1M—262K—6h agoIo Netzero-retention$0.27 in/1M$1.89 out/1M34 tok/s33Kfp815m agoMorphzero-retention$0.29 in/1M$2.40 out/1M25 tok/s131K—15m agoPhalazero-retention$0.32 in/1M$2.70 out/1M51 tok/s262K—15m agoSiliconFlowzero-retention$0.30 in/1M$3.20 out/1M34 tok/s262Kfp88h agoDeepInfrazero-retention$0.32 in/1M$3.20 out/1M59 tok/s262Kfp86h agoVenice AIzero-retention$0.33 in/1M$3.25 out/1M69 tok/s256Kfp815m agoCoreWeavezero-retention$0.60 in/1M$3.60 out/1M92 tok/s262Kfp815m ago
DeepInfra
Novita AI
MorphFull specs, hardware verdicts and benchmarks on theQwen3.6 27B model page →