Provider guide
Where to run Qwen3 8B
3 live listings tracked — output prices vary 3.3× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
13 tok/s
measured throughput · $0.12 input / $0.46 output per million tokens
Novita AI$0.035 in/1M$0.14 out/1M—128K—21 days agoOpenRouter$0.12 in/1M$0.46 out/1M—131K—30 min agoAlibaba CloudCHEAPEST$0.12 in/1M$0.46 out/1M13 tok/s131K—28 min ago
Novita AI
OpenRouterFull specs, hardware verdicts and benchmarks on the Qwen3 8B model page →