GPT API at 80% Off: Use GPT-4o via aiapi.cheap

Why This Post Exists

GPT-4o is great. The bill at the end of the month is not.

A chatbot doing 5,000 requests a day on GPT-4o (500 input, 1K output tokens each) racks up roughly $1,800/month at official prices. The same workload through aiapi.cheap on the Pro plan costs around $360/month. Same model, same response, 80% off.

This post shows the setup in plain code — no marketing fluff.

The Setup, In One Picture

from openai import OpenAI

client = OpenAI(
    api_key="sk-aic-YOUR_API_KEY",          # ← from aiapi.cheap dashboard
    base_url="https://aiapi.cheap/api/proxy/v1",  # ← swap this URL
)

resp = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

That's the full migration. Two lines change. Existing OpenAI SDK code keeps working.

Models Available

| Model | Best For | Pro Plan Pricing (input / output per 1M tokens) |

|---|---|---|

| gpt-4o | General-purpose, vision, function calling | $0.50 / $2.00 |

| gpt-4o-mini | Cheap fast tasks, classification | $0.03 / $0.12 |

For reference, official GPT-4o is $2.50/$10.00 per 1M tokens. GPT-4o mini is $0.15/$0.60. Pro plan is 80% off both.

Node.js / TypeScript

Same SDK, different language:

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.AIAPI_KEY!,  // sk-aic-...
  baseURL: "https://aiapi.cheap/api/proxy/v1",
});

const resp = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Write a haiku." }],
});

console.log(resp.choices[0].message.content);

Streaming

GPT-4o supports SSE streaming. Set stream: true and iterate chunks:

stream = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Stream me a poem."}],
    stream=True,
)

for chunk in stream:
    delta = chunk.choices[0].delta.content
    if delta:
        print(delta, end="", flush=True)

No difference from calling OpenAI direct. Same chunk shape, same finish reasons.

Function Calling / Tools

GPT-4o function calling works unchanged. Pass tools=[...] and tool_choice like normal:

resp = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "What's the weather in Tokyo?"}],
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get current weather for a city",
            "parameters": {
                "type": "object",
                "properties": {"city": {"type": "string"}},
                "required": ["city"],
            },
        },
    }],
)

Vision

GPT-4o vision works the same way — pass image URLs or base64-encoded images in the message content array. The proxy forwards binary content untouched.

Common Mistakes

Why You'd Use This Instead of OpenAI Direct

Three honest reasons:

1. 80% cheaper. That's the whole pitch.

2. One key for 5 vendors. Same SDK call, swap model to claude-sonnet-4-6 or gemini-2.5-pro and you've changed providers. No second account, no second credit card.

3. Crypto top-up. $5 minimum, USDT/BTC/ETH/100+ coins via Oxapay. No subscription.

If you only ever use GPT and you have a stable corporate billing relationship with OpenAI, stay direct. If you're a solo founder or building a side project and the bill stings, this is the cheapest legitimate way to use GPT-4o that we know of.

Next Steps

Get a key, swap two lines, ship something.