People ask us this constantly: which AI model should we actually be using? Usually what they mean is they picked one model, use it for everything, and something feels off. The writing is either too slow and expensive for simple stuff, or too shallow for the hard stuff. That mismatch is the actual problem, and it comes down to picking the right tool for the job instead of defaulting to whatever you opened first.
Opus, Sonnet, and Fable (from Anthropic's Claude family and the Fable storytelling tool) aren't interchangeable. They're built for different jobs, cost different amounts, and run at different speeds. Once you know what each one is actually good at, the choice gets easy.
When to Use Opus
Opus is the heavyweight. It's the model you reach for when the task actually requires reasoning: pulling apart a messy contract, working through a multi-step business problem, writing strategy documents where the logic has to hold up under scrutiny, or handling something with real financial or legal weight behind it.
The tradeoff is speed and cost. Opus takes longer and costs more per use than the lighter models. That's fine when the stakes justify it. It's a bad deal when you're using Opus to write a routine internal email or summarize a meeting that any competent model could handle in half the time.
Rule of thumb: if getting it wrong would actually hurt you, use Opus. If a mistake is a minor annoyance you'd just fix and move on, you don't need it.
When to Use Sonnet
Sonnet is the workhorse. It's fast, cheap relative to Opus, and good enough for the vast majority of day-to-day business writing and thinking tasks. Drafting emails, summarizing documents, answering customer questions, doing first passes on reports, brainstorming, cleaning up notes. This is where most of your team's AI usage should actually live.
The mistake we see a lot is businesses defaulting to the heaviest model out of a vague sense that more expensive means better output. For 80% of daily tasks, that's just burning time and money for no real gain in quality. Sonnet gets you there faster, and the output difference for routine work is often not noticeable.
Think of Sonnet as your default setting. Opus is what you escalate to when the task earns it.
When to Use Fable
Fable is a different animal entirely. It's built for storytelling and narrative work, not business analysis or reasoning tasks. If you're building out a brand story, scripting video content, writing narrative-driven marketing, or doing anything where the goal is to make people feel something rather than just understand a fact, Fable is the tool built for that job.
Don't try to make Opus or Sonnet do Fable's job. You can get passable narrative writing out of either, but it reads like exactly what it is: an analytical model doing its best impression of a storyteller. If narrative and emotional pull actually matter to the output, use the tool designed for it.
How to Actually Decide in the Moment
Here's the simple version we give clients:
Ask what the task actually needs. If it's complex reasoning with real stakes, Opus. If it's routine business writing or thinking, Sonnet. If it's narrative or storytelling, Fable. Most businesses can run on Sonnet for 80% of their volume, escalate to Opus for the tasks that genuinely need heavier reasoning, and bring in Fable specifically for narrative and brand storytelling work.
The businesses that get the most value out of AI aren't the ones using the fanciest model everywhere. They're the ones who've figured out which tool fits which task, and built that matching into how their team actually works day to day, so nobody has to think twice about it every time they open a new task.
If you want help figuring out which models fit which parts of your business and getting your team set up to use them without the guesswork, talk to us at Level Up AI. We'll walk through your actual workflows and tell you straight what belongs where.