Models

Konnect.ai provides access to the world's leading AI models. Use them individually or combine them for better results.

Overview

Each AI model has different strengths and characteristics. By combining multiple models, Konnect.ai helps you get more reliable and comprehensive answers.

Models Available

Providers

Max in Ensemble

OpenAI

GPT-5.2

gpt-5.2

Latest GPT-5 model with advanced reasoning capabilities.

Available

GPT-5.2 Chat

gpt-5.2-chat-latest

ChatGPT snapshot model optimized for conversational use cases.

Available

GPT-4o

gpt-4o

Most capable GPT-4 model. Best for complex reasoning and creative tasks.

Available

GPT-4 Turbo

gpt-4-turbo

Fast and capable with 128K context window.

Available

GPT-3.5 Turbo

gpt-3.5-turbo

Fast and cost-effective for simpler tasks.

Available

Anthropic

Claude 3 Opus

claude-3-opus

Most intelligent Claude model. Excellent for complex analysis.

Available

Claude 3 Sonnet

claude-3-sonnet

Balanced performance and speed. Great for most tasks.

Available

Claude 3 Haiku

claude-3-haiku

Fastest Claude model. Ideal for quick responses.

Available

Google

Gemini Pro

gemini-pro

Advanced reasoning with multimodal capabilities.

Available

Gemini Flash

gemini-flash

Fast and efficient for high-volume tasks.

Available

Coming Soon

Llama 3

llama-3

Meta's open-source model with strong performance.

Coming soon

Mistral Large

mistral-large

European AI with excellent multilingual support.

Coming soon

Choosing Models

When using Ensemble or Council modes, you can select which models to include. Consider these factors:

For accuracy

Combine GPT-5.2, Claude 3 Opus, and Gemini Pro for the most reliable results on complex questions.

For speed

Use faster models like GPT-3.5 Turbo, Claude 3 Haiku, and Gemini Flash when response time matters.

For diversity

Include at least one model from each provider to get diverse perspectives and reduce bias.

Model Notes

GPT-5.2 Chat

This model only supports temperature=1. Custom temperature values are automatically adjusted.

Aggregation Model

Ensemble synthesis and debate judgments use Claude Sonnet 4.5 by default for high-quality aggregation.

Using in API

Specify models in your API request using the models parameter:

JSON

{
  "message": "Explain quantum entanglement",
  "mode": "ensemble",
  "models": ["gpt-4o", "claude-3-opus", "gemini-pro"]
}