> ## Documentation Index
> Fetch the complete documentation index at: https://reedai-07fa30f1.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI Models

> Access GPT-4o, o1, o3, and other OpenAI models in ZeroTwo for advanced reasoning and generation

OpenAI provides some of the most advanced and widely-used language models available in ZeroTwo, from the versatile GPT-4o series to the specialized reasoning models like o1 and o3.

## Available models

<CardGroup cols={2}>
  <Card title="GPT-4o" icon="brain">
    **Context**: 128K tokens\
    **Strengths**: Balanced performance, multimodal, fast

    Best general-purpose model for most tasks
  </Card>

  <Card title="GPT-4o-mini" icon="zap">
    **Context**: 128K tokens\
    **Strengths**: Speed, cost-efficiency, large context

    Ideal for quick tasks and high-volume use
  </Card>

  <Card title="o1 / o3" icon="lightbulb">
    **Context**: 200K tokens\
    **Strengths**: Advanced reasoning, problem-solving

    Specialized for complex analysis and coding
  </Card>

  <Card title="GPT-4-turbo" icon="gauge-high">
    **Context**: 128K tokens\
    **Strengths**: Enhanced capabilities, vision

    Previous generation flagship model
  </Card>
</CardGroup>

## Model details

### GPT-4o (Omni)

OpenAI's flagship multimodal model with enhanced capabilities:

<ResponseField name="Context window" type="128,000 tokens">
  Large context for processing extensive documents and long conversations
</ResponseField>

<ResponseField name="Capabilities" type="features">
  * Text generation and analysis
  * Code generation and debugging
  * Image understanding (vision)
  * Function calling and tool use
  * JSON mode for structured outputs
  * Multimodal reasoning
</ResponseField>

<ResponseField name="Best for" type="use cases">
  * Complex coding tasks
  * Document analysis and summarization
  * Multimodal tasks (text + images)
  * API integrations with function calling
  * General-purpose chat and assistance
</ResponseField>

**When to use GPT-4o:**

* You need the best balance of capability and speed
* Working with images and text together
* Building applications with function calling
* Processing large documents
* Most production use cases

### GPT-4o-mini

Cost-effective model with impressive capabilities:

<ResponseField name="Context window" type="128,000 tokens">
  Same large context as GPT-4o
</ResponseField>

<ResponseField name="Capabilities" type="features">
  * Fast text generation
  * Code completion and basic debugging
  * Document processing
  * Function calling
  * JSON mode
  * Efficient for high-volume use
</ResponseField>

<ResponseField name="Best for" type="use cases">
  * Quick queries and simple tasks
  * High-frequency API calls
  * Chat applications
  * Content moderation
  * Data extraction and transformation
  * Cost-sensitive applications
</ResponseField>

**When to use GPT-4o-mini:**

* Speed is more important than maximum capability
* Processing many simple requests
* Budget constraints
* Real-time applications
* Simple code generation

### o1 and o3 (Reasoning models)

Advanced reasoning models for complex problem-solving:

<ResponseField name="Context window" type="200,000 tokens">
  Extended context for comprehensive analysis
</ResponseField>

<ResponseField name="Capabilities" type="features">
  * Extended reasoning and chain-of-thought
  * Complex problem decomposition
  * Advanced mathematical reasoning
  * Multi-step logical analysis
  * Code optimization and refactoring
  * Research and analysis tasks
</ResponseField>

<ResponseField name="Best for" type="use cases">
  * Complex algorithmic problems
  * Mathematical proofs and analysis
  * Advanced code architecture decisions
  * Research paper analysis
  * Multi-step reasoning tasks
  * Debugging complex issues
</ResponseField>

**When to use o1/o3:**

* Problem requires deep reasoning
* Multiple steps of logical thinking needed
* Complex code refactoring or optimization
* Mathematical or scientific analysis
* When accuracy is more important than speed

<Info>
  Reasoning models take longer to respond as they perform extended internal reasoning before generating output. Use them when the quality of reasoning justifies the wait.
</Info>

### GPT-4-turbo

Previous generation model, still highly capable:

<ResponseField name="Context window" type="128,000 tokens">
  Large context support
</ResponseField>

<ResponseField name="Capabilities" type="features">
  * Strong text and code generation
  * Vision capabilities
  * Function calling
  * JSON mode
  * Reliable performance
</ResponseField>

**When to use GPT-4-turbo:**

* Legacy applications
* Proven performance for specific use cases
* When GPT-4o is unavailable

## Capabilities comparison

| Capability           | GPT-4o | GPT-4o-mini | o1 / o3 | GPT-4-turbo |
| -------------------- | ------ | ----------- | ------- | ----------- |
| **Text generation**  | ⭐⭐⭐⭐⭐  | ⭐⭐⭐⭐        | ⭐⭐⭐⭐⭐   | ⭐⭐⭐⭐⭐       |
| **Code generation**  | ⭐⭐⭐⭐⭐  | ⭐⭐⭐⭐        | ⭐⭐⭐⭐⭐   | ⭐⭐⭐⭐        |
| **Reasoning**        | ⭐⭐⭐⭐   | ⭐⭐⭐         | ⭐⭐⭐⭐⭐   | ⭐⭐⭐⭐        |
| **Speed**            | ⭐⭐⭐⭐   | ⭐⭐⭐⭐⭐       | ⭐⭐      | ⭐⭐⭐         |
| **Vision**           | ✅      | ✅           | ❌       | ✅           |
| **Function calling** | ✅      | ✅           | ✅       | ✅           |
| **Cost efficiency**  | ⭐⭐⭐    | ⭐⭐⭐⭐⭐       | ⭐⭐      | ⭐⭐⭐         |

## Use cases by model

### GPT-4o use cases

<AccordionGroup>
  <Accordion title="Software development">
    * Full-stack application development
    * API design and implementation
    * Code review and optimization
    * Architecture planning
    * Debug complex issues
  </Accordion>

  <Accordion title="Content creation">
    * Blog posts and articles
    * Marketing copy
    * Technical documentation
    * Creative writing
    * Social media content
  </Accordion>

  <Accordion title="Data analysis">
    * Document summarization
    * Trend analysis
    * Report generation
    * Data extraction from text
    * Insight generation
  </Accordion>

  <Accordion title="Multimodal tasks">
    * Image description and analysis
    * OCR and text extraction from images
    * Visual content understanding
    * Screenshot analysis
    * Diagram interpretation
  </Accordion>
</AccordionGroup>

### GPT-4o-mini use cases

* Quick code completions
* Simple API integrations
* Chat and customer support
* Content moderation
* Simple data transformations
* Rapid prototyping
* High-volume processing

### o1 / o3 use cases

* Complex algorithm design
* Mathematical proofs
* Research paper analysis
* Advanced debugging
* System architecture design
* Security vulnerability analysis
* Complex refactoring

## Tips for using OpenAI models

<Steps>
  <Step title="Choose the right model">
    Start with GPT-4o for general use. Use GPT-4o-mini for simple tasks or when speed matters. Use o1/o3 for complex reasoning.
  </Step>

  <Step title="Leverage function calling">
    OpenAI models excel at function calling. Use this for API integrations and structured outputs.

    ```javascript theme={null}
    // Example: Function calling for structured data
    {
      "tools": [{
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get weather for a location",
          "parameters": {
            "type": "object",
            "properties": {
              "location": {"type": "string"},
              "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
            }
          }
        }
      }]
    }
    ```
  </Step>

  <Step title="Use JSON mode for structured output">
    When you need consistent JSON responses, enable JSON mode in your prompts.

    ```text theme={null}
    "Return the result as JSON with fields: title, summary, keyPoints"
    ```
  </Step>

  <Step title="Optimize context usage">
    * Place most important information early in prompts
    * Remove unnecessary context to save tokens
    * Use system messages for persistent instructions
  </Step>
</Steps>

<Tip>
  **Pro tips:**

  * GPT-4o works great with images—upload screenshots for debugging
  * o1/o3 models benefit from letting them "think"—don't interrupt reasoning
  * GPT-4o-mini is surprisingly capable for its speed and cost
  * Use temperature=0 for consistent, deterministic outputs
</Tip>

## Limitations and considerations

<Warning>
  **Be aware of:**

  **Knowledge cutoff**: Models have training data up to a specific date (varies by model). Use [web search](/tools/web-search) for current information.

  **Hallucinations**: Models may generate plausible but incorrect information. Verify critical facts.

  **Context limits**: Even with 128K tokens, extremely long contexts may degrade performance.

  **Reasoning model latency**: o1/o3 are slower due to extended reasoning—not suitable for real-time use.

  **Vision limitations**: Image understanding is advanced but not perfect—verify critical visual information.
</Warning>

## Pricing tiers

<Info>
  OpenAI models are available on ZeroTwo Pro and Enterprise plans. Model access and rate limits vary by subscription tier.

  * **Free tier**: Limited access to GPT-4o-mini
  * **Pro tier**: Full access to GPT-4o, GPT-4o-mini, o1
  * **Enterprise tier**: Enhanced limits, priority access, o3 access

  See [subscriptions](/account/subscriptions) for detailed pricing.
</Info>

## Switching models

In ZeroTwo, you can switch between OpenAI models mid-conversation:

<Steps>
  <Step title="Open model selector">
    Click the model name in the chat header or use the keyboard shortcut.
  </Step>

  <Step title="Select OpenAI model">
    Choose from available OpenAI models based on your subscription.
  </Step>

  <Step title="Continue conversation">
    The conversation continues with the new model, maintaining context.
  </Step>
</Steps>

Learn more about [switching models](/getting-started/switch-models).

## Related resources

<CardGroup cols={2}>
  <Card title="Models overview" icon="brain" href="/overview/models-and-providers">
    Compare all available models
  </Card>

  <Card title="Anthropic (Claude)" icon="sparkles" href="/providers/anthropic">
    Alternative high-capability models
  </Card>

  <Card title="Prompting guide" icon="message" href="/prompts/overview">
    Optimize prompts for OpenAI models
  </Card>

  <Card title="Function calling" icon="function" href="/tools/custom-extensions">
    Build tools with function calling
  </Card>
</CardGroup>