Air API is Now Live

Apr 9
2 min read

Updated: Apr 14

If you've ever tried serving an open-source AI model yourself, you know the pain. Setting up GPU infrastructure takes longer than choosing the model itself. Provisioning GPUs, configuring environments, scaling with traffic... the road to running a single model is way too long.

Air API eliminates that entire process. It's a serverless API service for open-source AI models. No infrastructure to build. Just an API key to get started.

Key Features

💡 OpenAI-Compatible Endpoint If you're already using the OpenAI API, just swap the endpoint URL. Almost no code changes required.

💡 Pay-Per-Token Pricing No monthly fixed fees. You only pay for the tokens you use.

💡 No Infrastructure Management GPU provisioning, model deployment, and scaling are all handled by AIEEV's distributed GPU infrastructure.

Launch Models

Air API currently supports 3 models from the Qwen series.

모델 가격 비교 표 한눈에 보기 > https://ap-1.aieev.cloud:3007/models

Model	Highlights	Recommended Use
Qwen3.5-35B-A3B	MoE architecture. Only 3B of 35B params active. 262K context.	AI agents, coding assistants, long-document analysis
Qwen3.5-9B	Dense model. Outperforms models 13x its size on benchmarks.	Real-time chatbots, cost-sensitive services
Qwen3-TTS (Beta)	Voice cloning from a 3-second sample. 10 languages including Korean.	Voice AI, multilingual guidance systems