Provider guide
Where to run GPT-5.6 Luna
8 live listings tracked — output prices vary 5.0× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
84 tok/s
measured throughput · $0.22 input / $1.32 output per million tokens
OpenRouterCHEAPEST$0.20 in/1M$1.20 out/1M—1.1M—2 hours agoMicrosoft Azure AIzero-retention$0.22 in/1M$1.32 out/1M66 tok/s1.1M—2 hours agoAmazon Bedrock$0.22 in/1M$1.32 out/1M84 tok/s1.1M—2 hours agoMicrosoft Azure AIzero-retention$0.22 in/1M$1.32 out/1M26 tok/s1.1M—2 hours agoOpenAI$1.00 in/1M$6.00 out/1M—1.1M—2 months agoMicrosoft Azure AIzero-retention$1.00 in/1M$6.00 out/1M—1.1M—52 days agoOpenAIflex tier$0.10 in/1M$0.60 out/1M48 tok/s1.1M—2 hours agoOpenAIfast tier$0.40 in/1M$2.40 out/1M38 tok/s1.1M—2 hours ago
Microsoft Azure AIFull specs, hardware verdicts and benchmarks on the GPT-5.6 Luna model page →