Provider guide
Where to run Granite 4.2 8B
3 live listings tracked — output prices vary 1.7× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
58 tok/s
measured throughput · $0.10 input / $0.15 output per million tokens
CoreWeavezero-retention$0.10 in/1M$0.15 out/1M58 tok/s131Kbf163 hours agoOpenRouterCHEAPEST$0.060 in/1M$0.25 out/1M—131K—3 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.060 in/1M$0.25 out/1M25 tok/sthrough OpenRouter131Kbf163 hours ago
DeepInfraFull specs, hardware verdicts and benchmarks on the Granite 4.2 8B model page →