Models

A guide to understanding and choosing the right AI model for your Mingleego assistant

Overview

This guide introduces the AI models available on the Mingleego platform, powered by OpenAI. These models drive your assistant’s capabilities, enabling tasks like answering questions, scheduling appointments, or generating content. Each model offers unique strengths, from multimodal processing to advanced reasoning, tailored to different use cases. Use this guide to select the best model for your assistant, whether for social chatbots (e.g., Facebook Messenger, WhatsApp, Instagram, Telegram, Discord) or web chatbots on your website.

Note: This guide assumes you have an account on the Mingleego platform. If not, please sign up here before proceeding.

Model Comparison

The table below compares the available models, their capabilities, and best use cases to help you choose the right one for your assistant.

Model	Best For	Inputs	Outputs	Context Window	Notes
GPT-4.1	Complex tasks without advanced reasoning	Text, Images	Text	1,047,576 tokens	Versatile for general-purpose tasks like customer support.
GPT-4.1 Mini	Balanced power and performance	Text, Images	Text	1,047,576 tokens	Great starting point for most chatbot applications.
GPT-4.1 Nano	Speed and efficiency	Text, Images	Text	1,047,576 tokens	Fastest for simple tasks like summarization or FAQs.
GPT-4.5 Preview	Human-like responses, complex workflows	Text, Images	Text	1,047,576 tokens	Experimental, ideal for advanced chatbots and long conversations.
GPT-5	Coding and agentic tasks	Text, Images	Text	400,000 tokens	Best for coding, reasoning, and technical workflows.
GPT-5 Mini	Well-defined technical tasks	Text, Images	Text	400,000 tokens	Faster, efficient for precise coding and STEM tasks.
GPT-5 Nano	Summarization and classification	Text, Images	Text	400,000 tokens	Fastest for lightweight tasks like data processing.
GPT-5.1	Flagship coding & agentic tasks	Text, Images	Text, Images	400,000 tokens	Configurable reasoning effort, reasoning token support, 128K max output tokens.
GPT-5.2	Fast & powerful flagship model	Text, Images	Text, Images	400,000 tokens	Fast flagship model with 30% fewer hallucinations, optimized for speed and quality.
GPT-5.4	Flagship coding & agentic tasks	Text, Images	Text, Images	400,000 tokens	Configurable reasoning effort; strong coding, tool use and agentic workflows.
GPT-5.4 Mini	Well-defined technical tasks	Text, Images	Text, Images	400,000 tokens	Faster and cheaper for precise coding and STEM tasks.
GPT-5.4 Nano	Summarization and classification	Text, Images	Text, Images	400,000 tokens	Fastest for lightweight, high-volume tasks like data processing.
GPT-5.4 Pro	Deep reasoning, hardest problems	Text, Images	Text, Images	400,000 tokens	Extended reasoning for complex, high-stakes work.
GPT-5.5	Fast & capable flagship	Text, Images	Text, Images	400,000 tokens	Top quality and speed with fewer hallucinations.
GPT-5.5 Pro	Most powerful reasoning	Text, Images	Text, Images	400,000 tokens	Maximum reasoning depth for the most demanding tasks.
GPT-5.6 Luna	Fast, high-volume conversations	Text, Images	Text, Images	—	Lightest GPT-5.6 tier: quick answers to FAQs and simple requests at the lowest cost.
GPT-5.6 Terra	Balanced everyday assistant	Text, Images	Text, Images	—	Middle GPT-5.6 tier: balances quality, speed and cost for support and sales chats.
GPT-5.6 Sol	GPT-5.6 flagship	Text, Images	Text, Images	—	Strongest GPT-5.6 tier for complex questions and tool use.
GPT-6 Luna	Fast GPT-6 for high-volume chats	Text, Images	Text, Images	—	Lightest GPT-6 tier: GPT-6 quality for quick, high-volume conversations.
GPT-6 Sol	GPT-6 flagship	Text, Images	Text, Images	—	Strong reasoning and tool use for demanding assistants.
GPT-6 Astra	Long context, hardest problems	Text, Images	Text, Images	Long context	Most powerful GPT-6 tier for very long documents and conversations; the highest price per token.
GPT-6.1 Sol	Newest flagship	Text, Images	Text, Images	—	Latest flagship model on the platform.

Choosing the Right Model

Selecting the right model depends on your assistant’s purpose and workload:

For tasks with images: every model in the table accepts images as input. For high volume on a budget, pick a Luna or Mini model; for the best understanding, a Sol model.
For general-purpose chatbots: Use GPT-4.1 or GPT-4.1 Mini for versatile tasks like customer support, FAQs, or content generation. GPT-4.1 Nano is best for simple, high-speed tasks.
For the newest models: GPT-6.1 Sol is the latest flagship. Within the GPT-6 and GPT-5.6 lines, the tiers go from light to most powerful: Luna for fast, high-volume chats (GPT-6 Luna, GPT-5.6 Luna), Terra for a balance of quality and cost (GPT-5.6 Terra), Sol for flagship quality (GPT-6 Sol, GPT-5.6 Sol), and GPT-6 Astra for very long context and the hardest tasks.
For advanced reasoning or coding: Select GPT-5.5 for a fast, capable flagship or GPT-5.5 Pro for maximum reasoning depth; GPT-5.4 for flagship coding and agentic tasks (with GPT-5.4 Mini, GPT-5.4 Nano and GPT-5.4 Pro variants); GPT-5.2 or GPT-5.1 for previous flagships, GPT-5 for complex workflows, GPT-5 Mini for precise technical tasks, or GPT-5 Nano for lightweight data processing.
For experimental or complex workflows: Try GPT-4.5 Preview for cutting-edge performance in long conversations or sophisticated applications.

Test your assistant’s performance with different models on the Playground page to find the best fit.

Next Steps

After choosing a model, you can:

Create an assistant: Follow the Creating an Assistant guide to set up your assistant with the selected model.
Configure integrations: Enhance your assistant with tools like Google Calendar using the Integrations guide.
Customize behavior: Define your assistant’s tone and responses with the System Instructions guide.

Start building your AI-powered assistant today with the model that best suits your needs!