> ## Documentation Index
> Fetch the complete documentation index at: https://prompt.university/llms.txt
> Use this file to discover all available pages before exploring further.

# GPT-4o

> OpenAI's flagship multimodal model with text, image, and audio capabilities

# GPT-4o

<Frame caption="OpenAI's flagship multimodal model">
  <img src="https://mintcdn.com/prompt-e71dd2c1/WKZkvyp6oa6K4aVy/images/banner.jpg?fit=max&auto=format&n=WKZkvyp6oa6K4aVy&q=85&s=7f8654ef1b7a176add1464ec7d2c6dbd" alt="GPT-4o Model Banner" width="2880" height="1620" data-path="images/banner.jpg" />
</Frame>

<Note>
  **Validation Status**: ✅ Validated by Graduate Layer\
  **Overall Rank**: #2 | **Confidence**: 95%
</Note>

## Overview

| Property           | Value              |
| ------------------ | ------------------ |
| **Provider**       | OpenAI             |
| **Release Date**   | May 13, 2024       |
| **Modalities**     | Text, Image, Audio |
| **Status**         | Validated          |
| **Context Window** | 128,000 tokens     |

## Capabilities

<AccordionGroup>
  <Accordion title="🎯 Core Capabilities" icon="star">
    * Chat and conversation
    * Image understanding and analysis
    * Audio understanding and generation
    * Function calling and tool use
    * Code generation
    * Reasoning and problem-solving
  </Accordion>

  <Accordion title="📊 Benchmark Scores" icon="chart">
    | Benchmark | Score | Rank |
    | --------- | ----- | ---- |
    | MMLU      | 88.7  | #3   |
    | HumanEval | 90.2  | #2   |
    | MATH      | 76.4  | #2   |
    | GPQA      | 53.6  | #2   |
    | MMMU      | 69.1  | #1   |
  </Accordion>

  <Accordion title="💰 Pricing" icon="dollar-sign">
    | Type   | Cost                |
    | ------ | ------------------- |
    | Input  | \$0.005 / 1K tokens |
    | Output | \$0.015 / 1K tokens |

    *45% cheaper than GPT-4 Turbo*
  </Accordion>
</AccordionGroup>

## Social Discovery

### X (Twitter) - @OpenAI

<Card>
  **Posted**: May 13, 2024\
  **Engagement**: 45K likes | 12K shares | 2.5M views | 8.9K bookmarks

  > "Introducing GPT-4o, our new flagship model that can reason across audio, vision, and text in real time."

  [View Post →](https://x.com/OpenAI/status/1789265359296908582)
</Card>

### LinkedIn - OpenAI

<Card>
  **Posted**: May 13, 2024\
  **Engagement**: 8.9K likes | 1.2K shares | 450K views | 2.3K bookmarks

  > "GPT-4o delivers GPT-4-level intelligence with faster processing and improved capabilities across text, vision, and audio."
</Card>

## Validation Notes

<Check>Multimodal across text, image, and audio - Verified</Check>
<Check>128k context window - Verified</Check>
<Check>Real-time audio processing (\< 300ms) - Verified</Check>
<Check>GPT-4 level intelligence - Verified</Check>
<Check>Superior vision capabilities (SOTA) - Verified</Check>
<Check>Improved function calling (94% accuracy) - Verified</Check>
<Warning>50% cost reduction claim - Partially verified (actual: 45%)</Warning>

## Use Cases

### Recommended For:

* Multimodal applications requiring vision + text
* Real-time audio interactions
* Cost-sensitive production deployments
* Vision-heavy tasks (analysis, description)

### Consider Alternatives When:

* Pure coding tasks → Claude 3 Opus
* Maximum reasoning required → Claude 3 Opus
* Lowest latency critical → GPT-4o-mini

## Related Models

<CardGroup cols={2}>
  <Card title="GPT-4o-mini" icon="zap" href="#">
    Faster, cheaper variant for simpler tasks
  </Card>

  <Card title="GPT-4 Turbo" icon="bot" href="#">
    Previous generation, higher cost
  </Card>
</CardGroup>

## Research Impact

<Update label="Jan 2026" description="Recent Paper Impact">
  [Chain of Thought Reasoning at Scale](/research/papers/2501-12345-reasoning-improvements) improved GPT-4o's mathematical reasoning by 15%
</Update>

## API Reference

```bash theme={null}
curl https://api.openai.com/v1/chat/completions \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

***

*Last validated: January 25, 2026 by researcher-alex*\
*First discovered: May 13, 2024*