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

AkashML

Official site · 🇺🇸 United States

Models it serves

6 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
gpt-oss-20bfp4$0.02$0.10131,07299.5%
gpt-oss-120bbf16$0.04$0.19131,072100.0%
Qwen3.6 35B A3Bfp8$0.10$0.90262,144100.0%
Qwen3.8-27Bfp8$0.20$1.78262,144100.0%
GLM-5.3fp8$1.05$3.561,048,576100.0%
Kimi K3fp4$3.00$15.001,048,576—