Model guide

Which AI model should I use?

You do not need deep technical knowledge to choose well. In Verics, each model is labeled with a plain-language category, so you can match it to your task in a few seconds.

Guide vs. Directory

This guide explains how to choose a model for your task. To see the models and providers currently available in Verics, browse the AI Model Directory.

The main decision factors

When you choose a model, you are really balancing a few simple factors against each other. You rarely need to optimise all of them at once — just the ones that matter for the task in front of you:

  • how complex the task is, from a quick rewrite to multi-step analysis
  • how much answer quality and depth you need
  • how quickly you want the response
  • whether the task involves files or images
  • whether it needs careful, step-by-step reasoning
  • whether it needs current or external information
  • how it fits your available usage

Model categories in Verics

Rather than asking you to memorise model names, Verics labels each model with a plain-language category. The specific models behind each category depend on your configuration, but the category tells you what to expect:

Fast

Best for simple questions, translations and short texts. Choose it when the task is straightforward, you want a quick draft, or you are rewriting or shortening existing text.

Balanced

A dependable default for most everyday tasks that are a bit more demanding. If you are not sure which category to pick, this is usually the right place to start.

Deep analysis

Best for complex, multi-step tasks and more extensive research. Choose it when the work involves several steps, a detailed comparison, or careful analysis where quality matters more than speed.

Match the model to the task

Beyond the category, a few task types are worth calling out:

Working with files or images

To have the AI read a document or picture, choose a model that supports it and attach a supported file. Whether a given model can work with images depends on the model, so pick one that supports the file type you are adding.

Current information

For anything that may have changed recently, combine a suitable model with web search — or a more thorough research option if it is available in your plan. Always check the sources for claims that matter.

Technical and coding tasks

A Balanced or Deep analysis model generally handles explaining code, spotting errors or drafting small snippets. A model specialised for technical work may also be available depending on your configuration.

Quick selection guide

Use this as a starting point, then adjust based on the result you get:

Your taskStart withWhy
Rewrite, shorten or translate textFastQuick and enough for simple text work
Draft an email, summary or short articleBalancedA good all-round default
Analyse, compare or plan across several stepsDeep analysisHandles complexity and detail
Work from a document or imageA model that supports files or imagesReads your attached material
Use recent or external informationBalanced or Deep analysis, with web searchLooks beyond built-in knowledge

If you are not sure

Start light, then step up

Begin with a fast, general-purpose model and switch to a deeper one only when the task clearly needs it. A heavier model is not automatically a better answer — it is slower and uses more of your available usage, so match the effort to the task.

Put it into practice

Open Verics, pick a category that matches your task, and see the difference for yourself.