Provider guide
Where to run Qwen3 VL 30B A3B Instruct
5 live listings tracked — output prices vary 1.9× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
45 tok/s
measured throughput · $0.13 input / $0.52 output per million tokens
Alibaba Cloud$0.13 in/1M$0.52 out/1M45 tok/s131K—5 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.15 in/1M$0.60 out/1M25 tok/sthrough OpenRouter262Kfp85 hours agoOpenRouterCHEAPEST$0.15 in/1M$0.60 out/1M—262K—5 hours agoNovita AIzero-retention through OpenRouterDirect and through OpenRouter$0.20 in/1M$0.70 out/1M39 tok/sthrough OpenRouter131Kbf165 hours agoSiliconFlowzero-retention$0.29 in/1M$1.00 out/1M14 tok/s262Kfp83 days ago
DeepInfra
Novita AIFull specs, hardware verdicts and benchmarks on the Qwen3 VL 30B A3B Instruct model page →