Cook with AI Enrollment closed
AI & LLMs Runs locally

Model Tier Quiz

Seven questions to pick budget, workhorse, or flagship LLM tier — with honest advice to start cheap.

ELI5

It helps you choose how powerful and expensive your AI model needs to be.

A little more detail

What this tool does

Many tasks do not need the most expensive model. Answer a few questions about difficulty, volume, latency, and failure cost. The result points you toward a budget, workhorse, or flagship tier.

Use it to
  • Pick a starting model tier for a new feature
  • Avoid paying flagship prices for routine work
  • Recognize tasks where better reasoning is worth the extra cost
Interactive workspace Results update as you type

Please answer every question before seeing your tier.

Your tier

Example models (mid-2026 snapshot)

Honest advice

Model tiers are a pricing ladder, not a personality test. Budget models handle bulk work; workhorses run most production apps; flagships are for when errors are expensive or the task needs the best reasoning available.

Scoring is rule-based — your answers add points to each tier. There is no AI behind this quiz. Example models are a static mid-2026 snapshot; check current pricing before you commit.

The real answer is almost always: start cheap, measure with evals on your actual prompts, and upgrade one tier when cheap models fail — not when marketing says you need the frontier model.