> ## Documentation Index
> Fetch the complete documentation index at: https://agent-docs.nineteen58.co.za/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Models

> Available language models and their capabilities

## Overview

Nineteen58 Agents provides access to a comprehensive selection of AI models from leading providers. This guide details each model's capabilities to help you choose the right one for your needs.

## OpenAI Models

### GPT-4 Series

#### GPT-4o

OpenAI's flagship model for mission-critical applications:

* Exceptional reasoning and analytical capabilities
* Highest accuracy among OpenAI models
* Best for complex problem-solving
* Ideal when precision is paramount

#### GPT-4o-Mini

Our recommended default choice for most applications:

* Balanced performance and efficiency
* Strong general-purpose capabilities
* Faster than GPT-4o
* Cost-effective for daily operations

#### GPT-4-Turbo

Optimized for real-time applications:

* Enhanced speed compared to standard GPT-4
* Recent knowledge base
* Strong performance in interactive scenarios
* Good for customer-facing applications

#### GPT-3.5-Turbo

Fast and efficient for straightforward tasks:

* Rapid response times
* Cost-effective
* Reliable for basic operations
* Perfect for simple automations

#### O1-Mini

Fast reasoning model for complex scenarios:

* Quick inference times
* Strong analytical capabilities
* Good for real-time complex tasks
* Balances speed and sophistication

## Anthropic Models

### Claude 3 Series

#### Claude-3-Opus

Anthropic's most advanced model:

* Superior analytical capabilities
* Excellent for research and academic work
* Detailed, well-structured responses
* Strong mathematical reasoning

#### Claude-3-Sonnet

Versatile model for business applications:

* Balanced performance profile
* Excellent content generation
* Reliable business communications
* Good cost-to-capability ratio

#### Claude-3-Haiku

Efficiency-focused model:

* Fast processing times
* Handles routine tasks well
* Cost-effective
* Ideal for high-volume operations

## Google Models

### Gemini Series

#### Gemini 2.0 Flash

Latest experimental model:

* Cutting-edge capabilities
* Optimized for real-time use
* Strong reasoning abilities
* Excellent interactive performance

#### Gemini 1.5 Pro

Robust general-purpose model:

* Comprehensive capabilities
* Reliable performance
* Good for complex tasks
* Strong business applications

#### Gemini 1.5 Flash

Speed-optimized variant:

* Very quick response times
* Good for real-time needs
* Maintains accuracy
* Perfect for interactive systems

### Gemma

Lightweight but capable:

* 7B parameter model
* Efficient processing
* Good for basic tasks
* Cost-effective solution

## Meta AI Models

### Llama Series

#### Llama (70B)

Large-scale open model:

* Comprehensive capabilities
* Good for complex tasks
* Fast processing on Groq
* Versatile applications

#### Llama (8B)

Efficient smaller model:

* Quick response times
* Good for simple tasks
* Cost-effective
* Lightweight deployment

## X AI Models

### GROK 2

Advanced conversational AI:

* Strong internet awareness
* Engaging personality
* Good for interactive applications
* Recent knowledge integration

### GROK 2 Vision

Multimodal capabilities:

* Image processing abilities
* Strong visual understanding
* Good for mixed media tasks
* Maintains GROK 2 capabilities

## Specialized Models

### Perplexity Sonar

#### Sonar Large

Powerful internet-aware model:

* Strong factual responses
* Recent information access
* Good for research
* Complex query handling

#### Sonar Small

Efficient internet-aware model:

* Quick response times
* Current information access
* Cost-effective
* Good for basic research

### Mixtral (8B)

Efficient performer:

* Strong performance for size
* Good speed-to-capability ratio
* Cost-effective
* Versatile applications

### Deepseek Models

#### Deepseek v3

Specialized reasoning model:

* Strong analytical capabilities
* Good for complex problems
* Detailed responses
* Research-oriented tasks

#### Deepseek r1

Capable conversational model:

* Good general-purpose performance
* Reliable responses
* Balanced capabilities
* Business-ready

## Choosing the Right Model

Consider these factors when selecting a model:

### Task Complexity

* **Simple Tasks**: GPT-3.5-Turbo, Claude-3-Haiku, Llama (8B)
* **Medium Complexity**: GPT-4o-Mini, Gemini 1.5 Pro, Sonar Small
* **Complex Tasks**: GPT-4o, Claude-3-Opus, GROK 2

### Response Speed

* **Ultra-Fast**: Gemini Flash models, GPT-3.5-Turbo
* **Balanced**: GPT-4o-Mini, Mixtral, Sonar Small
* **Accuracy-Focused**: GPT-4o, Claude-3-Opus, Deepseek v3

### Cost Efficiency

* **Budget-Friendly**: GPT-3.5-Turbo, Llama (8B), Gemma
* **Mid-Range**: GPT-4o-Mini, Claude-3-Sonnet, Mixtral
* **Premium**: GPT-4o, Claude-3-Opus, GROK 2

## Best Practices

1. **Start Conservative**: Begin with GPT-4o-Mini or Claude-3-Sonnet
2. **Benchmark Performance**: Test multiple models for your specific use case
3. **Monitor Metrics**: Track response times, costs, and accuracy
4. **Consider Hybrid Approaches**: Use different models for different tasks
5. **Regular Review**: Reassess as models are updated and new ones released

## Next Steps

* Learn to implement these models in the [Reasoning Builder](/docs/essentials/reasoning)
* Explore [Tool Integration](/docs/essentials/tools) capabilities
* Set up [Channel Integration](/docs/essentials/channels) for deployment