Provider guide
Where to run Gemma 4 26B A4B
14 live listings tracked — output prices vary 2.7× between the cheapest and most expensive host, so the provider choice matters as much as the model choice.
FASTEST MEASURED
74 tok/s
measured throughput · $0.10 input / $0.30 output per million tokens
Darkbloom$0.042 in/1M$0.22 out/1M41 tok/s131K—2 hours agoOpenRouterCHEAPEST$0.076 in/1M$0.26 out/1M—262K—2 hours agoNextBitzero-retention$0.076 in/1M$0.26 out/1M60 tok/s262Kbf162 hours agoCloudflare Workers AI$0.10 in/1M$0.30 out/1M49 tok/s256K—2 hours agoCoreWeavezero-retention$0.10 in/1M$0.30 out/1M74 tok/s262Kbf162 hours agoMakorazero-retention$0.080 in/1M$0.32 out/1M57 tok/s256K—2 hours agoDekaLLMzero-retention$0.060 in/1M$0.33 out/1M73 tok/s262Kbf162 hours agoDeepInfrazero-retention through OpenRouterDirect and through OpenRouter$0.070 in/1M$0.34 out/1M22 tok/sthrough OpenRouter262Kfp82 hours agoSiliconFlowzero-retention$0.14 in/1M$0.40 out/1M28 tok/s262Kfp82 hours agoVenice AIzero-retention$0.13 in/1M$0.40 out/1M26 tok/s256Kbf162 hours agoNovita AIzero-retention through OpenRouterDirect and through OpenRouter$0.13 in/1M$0.40 out/1M32 tok/sthrough OpenRouter262Kbf162 hours agoParasailzero-retention$0.13 in/1M$0.40 out/1M24 tok/s262Kbf162 hours agoIo Netzero-retention$0.15 in/1M$0.50 out/1M49 tok/s262Kbf168 hours agoGoogle Vertex AIzero-retention$0.15 in/1M$0.60 out/1M23 tok/s262K—14 hours ago
Cloudflare Workers AI
DeepInfra
Novita AI
Google Vertex AIFull specs, hardware verdicts and benchmarks on the Gemma 4 26B A4B model page →