BENCHMARKFirst benchmarks recorded: MiniMax M3 → Cortecs → ?6hBENCHMARKFirst benchmarks recorded: Kimi K3 → Cortecs → ?7hBENCHMARKFirst benchmarks recorded: DeepSeek V4 Pro → Cortecs → ?7hBENCHMARKFirst benchmarks recorded: GLM-4.7 → Cortecs → ?7hBENCHMARKFirst benchmarks recorded: GLM-5.2 → Cortecs → ?7h

Llama 3.3 70B

10 providers · 70B dense · 128k context · 2 regions · all models →

Llama 3.3 70B provider ranking

Updated 12m agoDEV DATA
#ProviderΔ30d
01FireworksBEST OVERALLUS · Serverless · Llama 3.3 70B99.0%98.8%99.9%2317901.8s251$0.12$0.48↑ 1.0
02NebiusBEST EUROPEANEU RESIDENTNL · Serverless · Llama 3.3 70B98.8%98.0%99.3%2137782.3s192$0.07$0.22↑ 1.4
03GroqUS · Serverless · Llama 3.3 70B98.5%98.1%98.6%1322980.5s1013$0.16$0.32↓ 0.5
04CerebrasFASTESTUS · Serverless · Llama 3.3 70B98.9%98.2%98.5%1102720.3s2835$0.33$0.66↑ 0.4
05OpenRouterBEST TOOLSUS · Routed · Llama 3.3 70B98.0%99.2%98.6%2849732.1s216$0.12$0.48↑ 0.5
06ScalewayEU RESIDENTEU OWNEDFR · Serverless · Llama 3.3 70B97.9%96.4%97.7%2639733.3s130$0.11$0.40↑ 1.3
07TogetherUS · Serverless · Llama 3.3 70B97.7%97.5%97.8%2979162.6s176$0.13$0.40↓ 0.7
08NextbitEU RESIDENTEU OWNEDES · Serverless · Llama 3.3 70B97.8%96.7%96.9%2528903.8s113$0.08$0.24↑ 1.8
10 of 10 providers ranked · Methodology →
Llama 3.3 70B inference benchmark: fastest & cheapest providers · InferenceBench