Provider guide
Where to run Qwen3.5-9B
7 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
56 tok/s
measured throughput · $0.10 input / $0.15 output per million tokens
Darkbloom$0.080 in/1M$0.13 out/1M18 tok/s262Kfp42 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.10 in/1M$0.15 out/1M8 tok/sthrough OpenRouter262Kbf168 hours ago direct2 hours ago through OpenRouterOpenRouterCHEAPEST$0.10 in/1M$0.15 out/1M—262K—2 hours agoVenice AIzero-retention$0.10 in/1M$0.15 out/1M56 tok/s256Kfp82 hours agoSiliconFlowzero-retention$0.10 in/1M$0.15 out/1M21 tok/s262Kfp82 hours agoParasailzero-retention$0.10 in/1M$0.25 out/1M26 tok/s262Kbf162 hours agoTogether AIzero-retention$0.17 in/1M$0.25 out/1M9 tok/s262K—2 hours ago
DeepInfra
Together AIFull specs, hardware verdicts and benchmarks on the Qwen3.5-9B model page →