How much does O4 cost?

O4 is billed pay-as-you-go at $0.45 per 1M input tokens and $1.83 per 1M output tokens. There is no subscription — you only pay for what you use.

What is the context window of O4?

O4 supports a context window of up to 200K tokens. It can return up to 100K tokens in a single response.

O4 is a paid, pay-as-you-go model — no subscription, you are only charged for usage.

How do I use O4 via the API?

O4 is OpenAI-compatible. Point any OpenAI SDK at https://api.airforce/v1 and pass the model ID o4 with your Api.Airforce API key.

O4 is OpenAI's chat model, served through the unified Api.Airforce gateway alongside 100+ other models.

OpenAIPaidOperational

O4

API model name: o4

O4 is OpenAI's chat model, served on the Api.Airforce unified API. It has a 200K-token context window. Beyond text, it accepts image as input. Capabilities include Vision, Tool calling, Reasoning, Documents, Prompt caching. It is priced at $0.45 per million input tokens and $1.83 per million output tokens. Knowledge cutoff: 2024-06. Access it through the OpenAI-compatible API with one key, alongside 100+ other models on Api.Airforce.

Get an API key View pricing

Pricing

Input / 1M tokens

$0.45

Output / 1M tokens

$1.83

Cache read / 1M tokens

$0.12

Specifications

Provider: OpenAI
Type: chat model
Context window: 200K tokens
Max output: 100K tokens
Knowledge cutoff: 2024-06
Input: text, image
Output: text
Prompt caching: Supported

Capabilities

VisionTool callingReasoningDocumentsPrompt cachingStreaming

Benchmarks

Independent evaluations and measured speed from Artificial Analysis.

Intelligence Index

25.6/100

Math Index

90.7/100

MMLU-Pro83%

GPQA Diamond78%

Humanity's Last Exam18%

LiveCodeBench86%

AIME 202591%

MATH-50099%

Output speed116.6 tok/s

Time to first token18.78 s

Source: Benchmark data by Artificial Analysis (artificialanalysis.ai)

What is O4 used for?

Chatbots & assistants — conversational AI, drafting, summarizing and Q&A.
Image understanding — analyze photos, screenshots, charts and scanned documents.
Agents & automation — function calling and tool use for multi-step workflows.
Complex reasoning — math, coding and step-by-step problem solving.
Document analysis — summarize and answer questions across long files.
Long-context tasks — process entire documents or codebases in a single prompt.
Real-time experiences — stream tokens for responsive chat and apps.

O4 vs. similar models

Model	Intelligence	Context	Input / 1M	Output / 1M
O4	25.6	200K	$0.45	$1.83
GPT 3.5	—	—	$0.47	$1.39
GPT 3.5 Turbo 16k	—	—	$10.26	$20.52
GPT 4	7.0	8K	$64.26	$128.52

Prices are Api.Airforce pay-as-you-go rates per 1M tokens. Context is the maximum input length.

O4 — frequently asked questions

How much does O4 cost?: O4 is billed pay-as-you-go at $0.45 per 1M input tokens and $1.83 per 1M output tokens. There is no subscription — you only pay for what you use.
What is the context window of O4?: O4 supports a context window of up to 200K tokens. It can return up to 100K tokens in a single response.
What can O4 do?: O4 supports Vision, Tool calling, Reasoning, Documents, Prompt caching.
Is O4 free to use?: O4 is a paid, pay-as-you-go model — no subscription, you are only charged for usage.
How do I use O4 via the API?: O4 is OpenAI-compatible. Point any OpenAI SDK at https://api.airforce/v1 and pass the model ID o4 with your Api.Airforce API key.
Who makes O4?: O4 is OpenAI's chat model, served through the unified Api.Airforce gateway alongside 100+ other models.

All models·Quickstart·Chat API reference

Use O4 via the API

OpenAI-compatible — point any OpenAI SDK at https://api.airforce/v1 and pass o4 as the model.

cURL

curl https://api.airforce/v1/chat/completions \
  -H "Authorization: Bearer $AIRFORCE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "o4",
    "messages": [{ "role": "user", "content": "Hello!" }]
  }'

Python

from openai import OpenAI
client = OpenAI(base_url="https://api.airforce/v1", api_key="$AIRFORCE_API_KEY")
r = client.chat.completions.create(
    model="o4",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(r.choices[0].message.content)

JavaScript

import OpenAI from "openai";
const client = new OpenAI({ baseURL: "https://api.airforce/v1", apiKey: process.env.AIRFORCE_API_KEY });
const r = await client.chat.completions.create({
  model: "o4",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(r.choices[0].message.content);

O4