Lesson 17 · Claude Mastery Pro+ ~10 min read Decision framework

Fable 5 vs Opus vs Sonnet: which Claude for which job now.

"Use the most powerful model" is bad advice. The most powerful model is slower, costs more, and is overkill for most of what you do all day. Here's a clear way to pick the right Claude for the task in front of you — and a plan for the billing that took over on July 20, 2026: Fable 5 included on Max, usage credits on Pro.

The mental model

Match the model to the task's difficulty, not your ambition.

Think of it like vehicles. You don't take a freight truck to pick up milk, and you don't move a houseful of furniture in a hatchback. Fable 5 is the freight truck: unmatched on long, heavy, complex hauls — and wasteful for a quick errand. Sonnet is the everyday car: fast, cheap, perfectly good for the trips you take most. Opus 5 (which replaced Opus 4.8 on July 24, 2026) sits in between — near-Fable intelligence at half the price, the new default on Max and the strongest model on Pro. Notably, the Opus tier is also what Fable quietly falls back to on sensitive topics.

01 The three Claudes at a glance

ModelBest atReach for it when…
Fable 5Long, complex, multi-step work; vision; tasks that run for a while on their ownThe job would take you hours and has many moving parts — a full analysis, a big migration, a research project, reading detailed figures
Opus 5Near-Fable reasoning and writing at half the price; default on Max, strongest on Pro (replaced Opus 4.8 on July 24, 2026)You want high quality but the task is bounded — a serious document, a tricky problem, careful reasoning that doesn't need hours of autonomy
SonnetSpeed and value on everyday tasksMost of your day: drafting emails, summarizing, quick questions, routine writing. Fast and inexpensive, and good enough that you won't miss the extra power
Reaching for Fable 5 on a 30-second task doesn't make the answer better — it just makes it slower and pricier. Power you don't need is waste, not insurance.

02 The 3-question decision

Before you pick, ask yourself three things:

  1. How long would this take me by hand? Minutes → Sonnet. An afternoon or more → Fable 5. In between → Opus.
  2. How many steps does it involve? One ask, one answer → Sonnet. A chain of dependent steps where each builds on the last → Fable 5.
  3. Does it involve images, charts, or documents I need read precisely? If yes, lean Fable 5 — vision is where its lead is largest right now.

Two or three answers pointing at "big" means Fable 5 earns its keep. Otherwise, save it — you'll get a faster answer and spend less.

03 Worked examples

Knowledge work

A useful habit

Start most tasks on Sonnet. If the answer feels thin, or the task turns out to have more moving parts than expected, then escalate to Fable 5. Escalating up is cheap; you rarely need to start at the top.

04 The money: pricing after the included window

Two pricing realities to plan around.

On the API, Fable 5 and Mythos 5 cost $10 per million input tokens and $50 per million output tokens — less than half what the earlier Mythos Preview cost. For developers, that's a meaningful price drop for top-tier capability.

On subscriptions, the promotional window closed July 19, 2026 — and the restore Anthropic promised happened, but only at the top. Since July 20, Fable 5 is a standard part of Max plans and premium Team and Enterprise seats: up to 50% of weekly limits at no extra cost, then usage credits. On Pro and standard Team or Enterprise seats it isn't included at all — it runs on pay-as-you-go usage credits, with a one-time free credit for eligible Pro and Team users.

How to spend credits wisely now

Pricing and access windows change fast — Anthropic has said it'll give notice before changes. Treat the dates here as accurate as of this writing and check Claude's current plan details before relying on them for a big workload.

Final challenge: sort your week

List the five Claude tasks you did this week. Tag each one Sonnet, Opus, or Fable 5 using the 3-question test. You'll likely find four are everyday Sonnet jobs and one is the real heavyweight. That ratio — mostly fast and cheap, occasionally powerful — is how to use the lineup well and keep your credits for what matters.

What you can do now

  • Match a task to the right Claude by difficulty, not by ambition
  • Use the 3-question test (length, steps, vision) to choose quickly
  • Default to Sonnet and escalate to Fable 5 only when the task is genuinely big
  • Understand the API pricing and how Fable 5 is billed on each plan since July 20, 2026
  • Spend usage credits on the tasks that actually justify a Mythos-class model
Pro
Up next in Claude Mastery

Lesson 18 · Fable 5's safeguards and your data: what actually changed

Why some requests get answered by an Opus model instead, why harmless ones occasionally trip the filter, and what the new 30-day data retention means for your company's AI policy. See pricing →