I want to talk about something that's quietly bleeding money out of a lot of businesses right now.
​
And the frustrating part is the math is sitting right there in plain sight.
​
Here's the situation.
​
Anthropic ships two main Claude models. Fable 5 at $10 per million tokens. Opus 5 at $5.
​
Most people assume more expensive equals more powerful. That's how everything else works. You pay more, you get more.
​
So they default to Fable 5 for everything.
​
Emails, Proposals, Research, and Daily work tasks.
​
Here's what the actual benchmarks show: Opus 5 scores higher on knowledge work. The cheaper model outperforms on the everyday stuff that makes up 90% of what most of us use AI for.
​
So a lot of people are paying double, getting less, and have no idea it's happening.
​
​Let me show you the math quickly.​
​
Say you're running 200 AI tasks a week. Mix of emails, summaries, content drafts, research pulls.
​
On Fable 5: $10 per million tokens. Average task, maybe 2,000 tokens. That's $0.02 per task. 200 tasks = $4/week, $208/year.
​
On Opus 5: $5 per million tokens. Same tasks. Same output. $0.01 per task. $104/year.
​
That's $104 back in your pocket for doing literally nothing different except switching the model.
​
Now scale that across a team of 5 people doing AI-assisted work.
​
$520 a year. Gone. For the worse model.
​
And that's before you factor in the users who crank up the "effort" setting on Fable 5. Effort multiplies tokens. Fable's tokens already cost 2x. That's double of double.
​
​Here's where it actually hurts.​
​
The money isn't even the real problem.
​
If Fable 5 just cost more and delivered more, fine. But it doesn't on most everyday tasks. So people are paying a premium and getting results that Opus 5 would have matched or beaten.
​
That means every email you wrote, every client proposal, every research summary where you defaulted to the expensive model... you paid double for the same answer.
​
That stings a little when you sit with it.
​
​So when does Fable 5 actually earn its spot?​
​
Fable for goals, Opus for tasks.
​
Fable 5 is built for the 20% of work that requires real depth. Complex contracts. A decision you only get one shot at. A problem where you need the model to think 10 moves ahead, not just draft the email.
​
Here's his actual framework:
​
Opus 5 prompt: "A client I've had two years went quiet after my renewal quote. I want to keep them. Ask me what you need to know before you suggest anything."
​
Fable 5 prompt: "Client of two years went quiet after my renewal quote. Thread attached. Goal: renewed at max 10% off, relationship intact. Work it end-to-end. Decide the approach, write the email, write the two follow-ups if they don't reply. Tell me how this backfires before you show me the draft."
​
See the difference?
​
Opus 5 has a conversation with you. You guide it. You edit. You send.

Fable 5 runs the whole operation. You hand it a goal and a context file. It comes back with a decision and a battle plan.
​
That's what the extra cost buys. Strategic depth on high-stakes, one-shot work. Not daily emails.
​
​The fix takes 30 seconds.​
​
Go into your Claude settings.
​
Set Opus 5 as your default.
​
Open Fable 5 when the work is complex enough to actually justify it.
​
That's it.
​
I've been doing this for a few weeks now. The outputs on routine work are either the same or better. The savings aren't dramatic on a personal level but across a business they add up fast.
​
And honestly, knowing I'm using the right model for the right job just feels better.
​
Our AiOS Certified Operators are trained on this kind of thing. Knowing which tool goes where, and positioning yourself as the person who does.

The business owners I talk to aren't spending time figuring out Claude's pricing tiers. That's the gap we fill.