Fine-Tune vs Prompt Engineering Calculator
Find the break-even point between fine-tuning and few-shot prompts.
With the default inputs, Break-even at 222,222 requests — Few-shot examples add $0.000225/request; fine-tuning removes them. Enter your own numbers below to recompute instantly; the full step-by-step math is shown under the worked example.
Results update automatically as you type.
- Few-shot tokens / request
- 1,500 tokens
- Added cost / request
- $0.000225
- One-time fine-tune cost
- $50.00
- Break-even volume
- 222,222 req
- Few-shot cost / month
- $45.00
Few-shot examples add tokens to every request forever; fine-tuning is a one-time cost that removes them. See how many requests it takes for fine-tuning to pay for itself.
How this is calculated
Few-shot prompting adds example tokens to every request forever; fine-tuning is a one-time cost that removes them. Break-even requests = one-time fine-tune cost ÷ (example tokens × input price per request). Above that volume, fine-tuning is cheaper.
Is this a good result? What to do next
If you send the same examples on high volume, fine-tuning usually wins on cost above the break-even. But it's not only about cost: prompting iterates faster, while fine-tuning can improve reliability and latency at the price of flexibility.
Typical planning ranges
- Break-even falls as
- volume and example size rise
- Low volume
- prompting usually wins
- High steady volume
- fine-tuning usually wins
Ranges are typical planning figures to sanity-check your result, not authoritative benchmarks. Your numbers will vary with use case, volume, and vendor.
How to improve this number
- Estimate real request volume as accurately as possible.
- Count all recurring example tokens, not just the average.
- Factor the (often higher) fine-tuned inference rate.
Common mistakes
- Ignoring that fine-tuned inference may cost more per token.
- Deciding on cost alone, not quality and flexibility.
When to use a different approach
To size the one-time training run, use the fine-tuning cost calculator. For the recurring example cost, use the few-shot prompt cost calculator.
Worked example (defaults)
With the default inputs above, here is the result:
- Few-shot tokens / request
- 1,500 tokens
- Added cost / request
- $0.000225
- One-time fine-tune cost
- $50.00
- Break-even volume
- 222,222 req
- Few-shot cost / month
- $45.00
- break-even = fine-tune cost / example tokens cost per request
Sources & references
Frequently asked questions
When is fine-tuning worth it?+
When you send the same examples on high request volume. Above the break-even volume, removing those tokens saves more than the training cost.
Are there non-cost reasons to choose one?+
Yes — prompting is faster to iterate; fine-tuning can improve reliability and latency but is less flexible.
Related calculators
Answering a real question?
This calculator powers these problem-solving guides: