Which Claude Model Should You Use? (Most of You Don't Need Fable)
An honest, cost-first guide to choosing a Claude model: what each costs, what the July 7 Fable change means, and why most people don't need the expensive model — from an agency that runs Claude in production daily.
Here's the honest version most guides won't give you: almost nobody needs Fable 5. It's a superb model — and it's overqualified for the work the average knowledge worker actually does all day. Default to Sonnet 5 ($3/$15 per million tokens, intro $2/$10 through August), drop to Haiku 4.5 ($1/$5) for high-volume simple work, escalate to Opus 4.8 ($5/$25) for genuinely hard reasoning, and reserve Fable 5 ($10/$50) for the narrow band that truly needs it: deep strategy, serious code and security analysis, and frontier long-context work. The price spread is why the distinction matters — Fable costs ten times what Haiku does — and as of today, July 7, 2026, it matters more than yesterday: the subsidized Fable access bundled with paid plans just ended, so casual Fable use is now billed at that top rate. We run production systems on Claude every day and keep the bill boring by routing each job to the smallest model that can do it. Here's the map, the math, the three levers that cut cost without downgrading the work, and an honest take on who actually needs the expensive model.
Every "which Claude model" guide online tells you what each model is good at. Almost none tell you what each one costs you, whether you're overpaying, or — the part nobody says out loud — that most people reaching for the top model don't need it. This is that version, from people who watch the meter.
What each Claude model costs
Claude has four models, and the price gap between the cheapest and the most expensive is 10×, which makes "just use the best one" a quietly expensive habit. Here are the current API rates, per million tokens, verified on the publish date. These are also the rates you pay once you exceed a subscription plan's limits and switch on usage credits — so they matter even if you never touch the raw API.
| Model | Input / M tokens | Output / M tokens | Built for |
|---|---|---|---|
| Haiku 4.5 | $1 | $5 | Fast, high-volume, mechanical work |
| Sonnet 5 | $3 ($2 intro*) | $15 ($10 intro*) | The balanced daily driver |
| Opus 4.8 | $5 | $25 | Deep reasoning, hard coding, high-stakes review |
| Fable 5 | $10 | $50 | Frontier autonomy, very long context and output |
*Sonnet 5 launched at an introductory $2/$10 per million tokens through August 31, 2026, then moves to its standard $3/$15.
To make the spread real: a month of 100 million input tokens and 50 million output tokens runs about $1,050 on Sonnet, $1,750 on Opus, and $3,500 on Fable. Same work, same volume — the model choice alone more than triples the bill. That single fact is the whole of model economics, and it's why the smartest teams treat model selection as a cost decision, not just a capability one.
The Fable question: what changed today
As of July 7, 2026, Claude Fable 5 is billed as metered usage credits at $10 input / $50 output per million tokens — roughly double Opus and ten times Haiku — now that its subsidized period on paid plans has ended. Through today, Pro, Max, Team, and select Enterprise plans included Fable for up to about half of your weekly usage limit. Starting now, that included allowance is gone; if you keep using Fable, you're paying API rates on top of your subscription.
The part that catches people: if you haven't enabled usage credits in the Claude Console, Fable access simply stops — no silent bill, but no access either. And if you have enabled credits without a spending cap, a few long Fable sessions on a big context can quietly dwarf your monthly subscription, because it's the most expensive model Anthropic sells. Nobody overspends here on purpose. They overspend by leaving "use the newest, best model" as the default and not noticing the meter turned on. The two-minute fix is to open the Console, decide whether you even need Fable enabled, and set a monthly cap before you forget.
The honest take: most of you don't need Fable
Fable 5 is an excellent model, and most people reaching for it are wrong to — it's too smart for the work the average knowledge worker actually does. That's not a knock on the model; it's a knock on the reflex. "Use the most capable model available" feels responsible and is usually just expensive. The daily work of most knowledge workers — drafting, summarizing, answering, analyzing a normal-sized document, writing and reviewing ordinary code — sits comfortably inside what Sonnet already does well, faster and at a fraction of the price. Paying ten times more for headroom you never touch isn't buying quality; it's buying idle horsepower.
So where does Fable actually earn its rate? In our experience, a narrow and real band. Deep strategy work, where you want a model holding a lot of moving context and reasoning several steps ahead before it commits. Serious code and security analysis, where a missed edge case is expensive and the extra reasoning depth pays for itself. And genuinely long-context, long-running autonomous tasks that need its very large context window and extended output. That's roughly the list. Those jobs are worth the top rate because the stakes or the sheer context justify it. Notice what isn't on it: most of what most people do most of the day.
Said plainly, the contrarian rule is this: reach for Fable only when the task is hard enough, or the stakes high enough, to justify a 10× premium — and default to Sonnet for everything else, which is nearly everything. Which reframes today's news. The Fable credits expiring on July 7 are, for the overwhelming majority of users, a non-event. If it genuinely stings, you're probably one of the few using Fable for what it's actually for. If you're not sure whether you're in that group, you're almost certainly not — and Sonnet has you covered.
Which model for which job
Think of the four models as a staff with four pay grades, and assign work by difficulty, not prestige: default to Sonnet, drop to Haiku for bulk, escalate to Opus for genuinely hard problems, and reserve Fable for the rare frontier task. Here's the mapping in practice.
Haiku 4.5 — the bulk worker. Classification, tagging, extraction, format conversion, routing decisions, first-pass triage, simple boilerplate. Anything high-volume and rule-shaped. At a fifth of Sonnet's price and a tenth of Fable's, Haiku is where cost-conscious throughput lives — and a common pattern is to run Haiku first on incoming work and only escalate the items it flags as hard.
Sonnet 5 — the default. This is what most work should run on: the majority of coding, content drafting, data analysis, and everyday reasoning. It outperforms earlier Opus models on many benchmarks while running much faster and cheaper, and its introductory pricing through August makes it an especially easy default right now. There's a genuine tension worth knowing about here: Anthropic's own model-overview docs suggest starting with Opus, while the practitioner community and Claude's own selection guide say start with Sonnet. The cost lens breaks the tie — Opus is more than twice Sonnet's output price, so Sonnet-first with deliberate escalation is the cost-correct default for all but the hardest work.
Opus 4.8 — the specialist. Reach for Opus when a task needs deeper reasoning, careful architectural planning, intricate debugging, or high-stakes review that Sonnet visibly fumbles. Crucially, it's half the price of Fable, so it's the right "step up" for most hard problems — you rarely need to jump straight to the top tier.
Fable 5 — the frontier tool, not the default. Reserve it for genuinely complex, multi-step autonomous work, or when you specifically need its very large context window or long maximum output. A powerful pattern: let the expensive model plan and review while a cheaper one does the volume execution — Fable or Opus scopes the work and checks the result, Sonnet or Haiku does the middle. You pay frontier prices only for the thinking, not the typing.
Three levers that cut the bill without downgrading the work
Beyond picking the right model, three settings cut model cost dramatically: prompt caching, the Batch API, and disciplined routing. We use all three, and together they're why an agency running autonomous systems all day pays predictable, boring bills.
| Lever | What it does | Use it when |
|---|---|---|
| Prompt caching | Cuts repeated-context input cost by ~90% (Fable cached input drops from $10 to ~$1/M) | You re-send the same system prompt, tool definitions, or file context every turn — i.e. any long session or loop |
| Batch API | Halves both input and output rates (Fable → Opus-priced, Sonnet → near-Haiku) | The work doesn't need to happen in real time: overnight analysis, bulk generation, scheduled reports |
| Model routing | Runs each task on the smallest capable model; reserves Fable for jobs that need its context/output | Always. This is the discipline that separates predictable bills from billing surprises |
Prompt caching is the highest-leverage of the three for anyone running long sessions, and it's essentially a switch. Batch pricing is free money for any workload nobody is watching in real time — which describes most business automation. And routing is the habit that makes the other two matter: cache and batch a Fable workload that should have been a Sonnet workload, and you've still overpaid.
What we actually run
Our five production loops — including the one that runs this site's SEO every day — cost about $200/month in fixed subscription plus a few dollars of data APIs, and a big reason is that they never run on Fable. Scheduled, bounded work on Sonnet- and Haiku-tier models is cheap and entirely good enough for the job: pull the search data, check the answer engines, draft the fixes, reconcile the books. We escalate a model only when a specific task genuinely earns it, and the frontier tier almost never does for recurring operational work.
That discipline is the same one behind the honest agency economics we've written about, and it's exactly why business loops stay affordable as you add more of them — the marginal cost of the next scheduled run on Sonnet is close to nothing. The teams that get surprised by an AI bill are almost always the ones defaulting to the biggest model for work the workhorse would have handled.
So which should you actually pick?
If you use Claude in the app, the model choice is mostly handled for you — just don't manually default everything to Fable; if you build on the API, route deliberately and set a cap. For a chat or Cowork user, stay on Sonnet for daily work and let the product escalate when it needs to; reserve manual Fable selection for the rare heavy task. For an API builder or anyone running automations, the routing map above is the difference between a predictable line item and a scary one — default to Sonnet, drop to Haiku for volume, escalate to Opus for hard problems, cache and batch what you can, and keep Fable on a leash with a Console spending cap.
And if your usage is heavy or spiky enough that you're weighing plans against the API, that's a related but separate decision — we broke down the subscription tiers (Pro, Max, Team, Enterprise) and when each one beats metered usage in our Claude plans breakdown. If you'd rather just get a read on where your own operation is overspending, that's what our Revenue Audit is for.
FAQ
Published: July 2026. All model rates, the Sonnet 5 introductory window, and the Fable 5 usage-credit change (effective July 7, 2026) were verified against Anthropic's pricing pages and documentation on the publish date. Claude pricing changes frequently — confirm current numbers before you buy or enable usage credits.
Related: Claude plans: Pro vs Max vs Team vs Enterprise · Claude Cowork pricing · The economics of an AI automation agency · Business loops · Claude loops for business