How many tokens does an epic take to build with AI?
A epic typically takes about 10,500,000 tokens to build with an AI coding agent (roughly 4,200,000–25,100,000 depending on complexity), about $30.00 at today's GPT-4o prices ($12.00–$71.00). This is the build-time (developer) cost, not the cost of running the finished feature.
An epic bundles many stories into one deliverable capability. Its AI build cost is essentially the sum of its stories, so it runs into tens of millions of tokens — the level at which prompt caching and model choice materially change the bill.
Token & cost ranges
Cost uses GPT-4o at current prices. Verified 29 July 2026.
| Scenario | Input tokens | Output tokens | Total tokens | Est. cost |
|---|---|---|---|---|
| Low (simple) | 4,000,000 | 200,000 | 4,200,000 | $12.00 |
| Typical | 10,000,000 | 500,000 | 10,500,000 | $30.00 |
| High (complex) | 24,000,000 | 1,100,000 | 25,100,000 | $71.00 |
Typical cost by model
The same typical token estimate priced across a few common coding models.
| Model | Typical cost | Range (low–high) |
|---|---|---|
| GPT-4o | $30.00 | $12.00 – $71.00 |
| GPT-4o mini | $1.80 | $0.7200 – $4.26 |
| Claude Sonnet 4.6 | $37.50 | $15.00 – $88.50 |
| Gemini 2.0 Flash | $1.20 | $0.4800 – $2.84 |
How we estimate this
An epic is a collection of stories delivering one capability — modelled here as roughly 8-15 stories of mixed size built over many sessions. Input tokens dominate because an agent re-sends file context on every turn; output is the generated code, diffs and tests. Treat these as planning ranges, not a quote — your real usage depends on codebase size, model, prompt caching and how many iterations it takes.
Examples of this size
- A complete billing & subscriptions capability
- An onboarding flow with several screens and rules
- A reporting module: data model, API, UI and tests
Estimate it with your own model
Plug in your model, complexity and prompt-caching to get a tailored range.
Frequently asked questions
How many tokens does a epic take to build with AI?+
A epic typically takes about 10,500,000 tokens to build with an AI coding agent (roughly 4,200,000–25,100,000 depending on complexity), about $30.00 at today's GPT-4o prices ($12.00–$71.00). This is the build-time (developer) cost, not the cost of running the finished feature.
Is this the cost to build it or to run it?+
This is the build-time (developer) cost — the tokens an AI coding agent burns while implementing the work. Running the finished feature is a separate, usually much smaller per-request cost.
Why is the range so wide?+
An epic is a collection of stories delivering one capability — modelled here as roughly 8-15 stories of mixed size built over many sessions. Input tokens dominate because an agent re-sends file context on every turn; output is the generated code, diffs and tests. Treat these as planning ranges, not a quote — your real usage depends on codebase size, model, prompt caching and how many iterations it takes.
How can I lower the cost?+
Split large work into smaller pieces, enable prompt caching so re-sent context is billed at the cheaper cache-read rate, and use a cheaper model for routine steps. Try your own model in the Cost to build X estimator.