Model Spend Arena2ND ED.
serves endpoints · 2026-10-03

GMICloud

Official site · 🇺🇸 United States

Models it serves

20 endpoints, cheapest first. This is the view the model-by-model tables cannot give you: what one provider actually offers, and at what precision.

ModelQuant$/M in$/M outContextUptime 30m
GLM-5.3-Flashfp8$0.09$0.301,048,57699.1%
Xiaomi MiMo-V2.5fp8$0.12$0.241,050,00098.2%
MiMo-V2.6-Flashbf16$0.13$0.271,050,00092.0%
DeepSeek V3.2fp8$0.21$0.31163,84091.5%
Tencent Hy3bf16$0.14$0.58262,144100.0%
DeepSeek V4.1 Flashfp8$0.18$0.721,048,57599.8%
MiniMax M2.7fp8$0.21$0.84196,60899.5%
Xiaomi MiMo-V2.5-Probf16$0.30$0.611,050,00092.2%
MiniMax-M3fp8$0.24$0.961,048,57699.2%
DeepSeek V4 Flashfp8$0.29$0.861,048,575100.0%
DeepSeek V3 0324fp8$0.29$1.14128,00099.1%
MiMo-V2.6-Probf16$0.41$0.831,050,00098.7%
DeepSeek V4 Flash Visionfp8$0.44$1.321,048,575100.0%
Qwen3.5 397B A17Bfp8$0.60$3.60262,14499.4%
GLM-5.3fp8$0.98$3.081,048,576—
Kimi K2.6fp8$0.85$3.60262,144—
DeepSeek V4 Profp8$1.06$3.171,048,575100.0%
Kimi K2.7 Codefp8$0.95$4.00262,144—
GLM-5.2fp8$1.40$4.401,048,576—
GLM-5.1fp8$1.40$4.40202,75299.0%