Using Fable 5: Effort Levels & Keeping Costs Low
Note: Fable 5 is the most capable model — and the priciest. The good news: you rarely need full power. This guide shows how to match the effort level to the job so you get great work without the big bill.
The tradeoff, in real numbers
The same site, built from the same prompt, at three effort levels:
| Effort | Time | Cost |
|---|---|---|
| Low | 1m 54s | $1.94 |
| Medium | 5m 54s | $7.49 |
| High | 13m 45s | $23.91 |
Same prompt, same site — a ~12× cost spread. The point: use the lowest effort that gets the job done.
5 moves to keep costs low
1. Match the effort level to the project
- Low — small, well-defined tasks
- Medium — real work you want polished
- High / full Fable — hard one-shots or builds worth it
2. Split big projects — plan expensive, execute cheap
- Use high effort for the planning, then switch to low effort to execute.
- Or hand execution to Opus for a small, defined task.
- Write less code, spend less: a ruleset like Ponytail makes the agent reuse what already exists instead of writing new code (their benchmarks: ~22% fewer tokens, ~20% lower cost).
3. Make low effort punch above its weight
- Low effort is already strong. If you have time, create more context and be more specific — it'll do great work cheaply.
4. Spend big only when it's worth it
- When you need one shot done well, or it's a project worth an overnight run, give Fable full effort and let it build from scratch.
5. Guardrails so Fable doesn't burn money
- Be very specific about the goal so it doesn't go down a wormhole.
- Start every task on a clean slate.
- Make it plan first — ask Fable what it'll do before you let it run.
Get the next guide when it's ready.
You're on the list.
No emails yet. Your address stays private.