Provider guide
Where to run Qwen3 30B A3B Instruct 2507
12 live offers tracked — output prices vary 2.8× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
161 tok/s
measured throughput · $0.12 input / $0.50 output per million tokens
StreamLake$0.048 in/1M$0.19 out/1M16 tok/s128K—42m agoOpenRouterCHEAPEST$0.048 in/1M$0.19 out/1M—262K—44m agoNovita AI$0.090 in/1M$0.45 out/1M—41K—7h agoDeepInfra$0.12 in/1M$0.50 out/1M—41Kfp87h agoAlibaba Cloud$0.13 in/1M$0.52 out/1M88 tok/s131Kfp842m agoNextBit$0.12 in/1M$0.52 out/1M5 tok/s33Kfp86d agoAlibaba Cloud$0.13 in/1M$0.52 out/1M68 tok/s131K—7d agoPhala$0.15 in/1M$0.55 out/1M69 tok/s262K—6d agoNebius AI Studiozero-retention$0.10 in/1M$0.30 out/1M28 tok/s262Kfp842m agoSiliconFlowzero-retention$0.090 in/1M$0.30 out/1M26 tok/s262Kfp838h agoCoreWeavezero-retention$0.10 in/1M$0.30 out/1M81 tok/s262Kbf1642m agoDeepInfrazero-retention$0.12 in/1M$0.50 out/1M161 tok/s41Kfp842m ago
Novita AI
Nebius AI StudioFull specs, hardware verdicts and benchmarks on theQwen3 30B A3B Instruct 2507 model page →