# Google Jules Pricing: Free, Pro and Ultra Plans (2026)
By Dario Espinoza-Blum · 2026-10-28 · Source: https://www.activepieces.com/blog/google-jules-pricing-free-pro-and-ultra-plans-2026
---
According to Google [Jules](https://jules.google/docs/usage-limits/), it starts free at 15 tasks per rolling 24 hours. Pro raises that to 100, and Ultra to 300, which means you'll exhaust the free tier within days, especially if you're also chaining it into automations built with tools like [Activepieces](https://www.activepieces.com).
Google Jules pricing is the set of free and paid subscription tiers for Google's asynchronous AI coding agent, which differ mainly in how many tasks you can run per rolling 24 hours and how many can run at once.
You hand it a job, such as fixing a bug, and it works in the background while you do something else. That is Jules, Google's asynchronous coding agent, and the cap on those jobs is what you're really buying.
### The plan table, checked 2026-10-28
The jump between tiers is much bigger than the price of entry suggests. The table shows what each tier gives you in daily tasks and in tasks running at once. Task figures come from [Google Jules](https://jules.google/docs/usage-limits/). The Pro price comes from [Plan-Price](https://plan-price.com/en/countries/united-states/), checked 2026-10-28.
| Plan | Monthly price | Daily tasks (rolling 24h) | Concurrent tasks | Inclusions |
|---|---|---|---|---|
| Jules Free | $0 | 15 | 3 | Jules at base limits |
| Jules in Pro | $19.99 (Google AI Pro, Plan-Price, checked 2026-10-28) | 100 | 15 | Jules bundled with the subscription |
| Jules in Ultra | Not confirmed (Google AI Ultra; check Google's plan page) | 300 | 60 | Jules bundled with the subscription |
| Google AI Plus | Not listed for Jules | Not listed | Not listed | No Jules entry |
_Prices and plan limits checked against [docs.claude.com](https://docs.claude.com/en/docs/about-claude/models/overview) and [claude.com](https://claude.com/pricing) and [openai.com](https://openai.com/chatgpt/pricing) and [gemini.google](https://gemini.google/subscriptions) on October 27, 2026._
Free allows 15 tasks a day, with 3 running at once, so your throughput is capped at three concurrent tasks.
On Pro, 100 a day and 15 at once mean a small team can queue a backlog without waiting. Ultra allows 300 a day and 60 at once, which only pays off if you run Jules like a production line.
I could not confirm a monthly price for Ultra from a source I could check on 2026-10-28, so I haven't put a figure in the table.
Pro to Ultra triples the daily cap from 100 to 300 tasks, so read Ultra's price on Google's page and ask whether the gap per extra task is worth it for your load.
### What counts as a task
Start a job and it counts against your cap for 24 hours from the moment you start it. The window rolls rather than resetting at midnight, so a burst at 9 a.m. leaves you short until those tasks age out the next morning.
A fourth job waits on Free even if you have daily tasks left, because concurrency is a separate limit: Free runs 3 jobs at a time.
### How Jules is bundled with Google AI plans
You don't buy Jules alone at the higher tiers. Pro and Ultra come as part of the matching Google AI subscription, so you're paying for the whole bundle and Jules is one piece of it.
Google AI Plus has no Jules entry, so don't assume it gives you anything beyond Free.
Before you pay, check each figure against Google's pages on the day you buy.
## What Jules costs for different builder types
Because Google can change its numbers, the useful way to price Jules is by how many tasks you run. A solo founder can stay free, a freelancer likely needs Pro, and a four-person team should test Ultra.
Task counts below are my estimates, not Google's figures.
Read the fit column first, then check the cap and price cells against Google's pages on the day you buy. The table shows each builder's estimated daily load against the plan that should cover it.
| Builder | Estimated tasks per day (my estimate) | Cap on Free / Pro / Ultra | Monthly cost | Fit |
|---|---|---|---|---|
| Solo founder | Well under one a day on average | Free is likely enough; Pro and Ultra are far more than needed | Nothing on Free | Free |
| Freelancer | A few a day, more in busy weeks | Free is the likely wall; Pro is the plan to test | Pro's listed price | Pro |
| Four-person team | Many at once, all day | Concurrency is the limit to watch; Ultra is the one to test | Ultra's listed price | Ultra, if Pro stalls them |
Look at your busiest days, not your average ones, because cost follows them. **Pick the plan that survives your worst week.**
If you hand Jules the odd bug fix or dependency update, you will rarely touch a daily cap.
Start on Free, and upgrade only after you hit the limit twice in a month, because one bad day isn't a pattern.
### Freelancer: daily tasks across client repos
When you switch between clients, tasks pile up on deadline days. A cap that looks generous on a quiet Tuesday can stop work on a Friday before an invoice goes out.
If a stalled task costs you a billable afternoon, **a paid plan is cheaper than the delay**. Count your heaviest recent week and compare it with Pro's cap.
### Four-person team: parallel tasks all day
Four people each starting a task will queue behind each other if the concurrency limit is low, and **a queue is lost time for four salaries**. The daily cap matters too for a team, but so does how many tasks can run at once.
Test with a real working day before committing, and confirm how Google counts people and tasks on its pricing page.
Starting those tasks by hand is also where a team spends cap on forgotten or duplicate clicks, and Activepieces can start them from a labelled ticket or pull request instead.
Its published pricing page at https://www.activepieces.com/pricing lists 1 credit per flow run however many steps the flow has, where Zapier's pricing page bills per task and Make's per module, so a long flow costs the same as a short one.
The free plan includes 1,000 credits a month and unlimited flows.
## Understand Jules limits and upgrade triggers
Count your tasks on a busy day before you count dollars. The limits that force an upgrade are the daily task cap and the concurrent task cap, not the monthly price. Five limits do the damage:
- The daily cap, counted over a rolling 24 hours, means a burst of work in the morning can lock you out that afternoon.
- The concurrent cap means you wait in line even when you have daily tasks left.
- Model access depends on your tier, so a cheaper plan may give you a weaker engine for the hard jobs.
- Limits shared with other Google AI features mean your chat use elsewhere can eat into your coding budget.
- Review time on generated pull requests (a pull request's a proposed code change waiting for your approval) is a cost in your hours, and no plan removes it.
Each one is covered below.
### Daily and concurrent task caps
Google has moved these numbers before. [TestingCatalog](https://www.testingcatalog.com/google-launches-coding-agent-jules-in-beta-with-free-daily-tasks/) reported 5 free tasks per day in May 2025, and 15 per day by August 2025, so users could accomplish three times as much for free by summer's end.
If you had chosen a plan on the old figure, you would have overbought. The reverse can happen too: a cap can shrink, and a workflow built around 15 would stall. Confirm both caps on Google's own pages on the day you pay.
Because a failed task on a weak model still uses up one of your daily tasks, it matters that higher tiers can unlock stronger models and bigger allowances.
If your work is mostly small fixes, the entry tier may be enough. If it's multi-file changes, the gated model is what you're really buying.
### Review time and failed tasks
Finishing a task doesn't mean you're done with it. Every pull request needs a person to read it, and a wrong one costs you the task, the reading time and the retry.
Say you run ten tasks a day and two come back unusable: you've spent the cap on work you must repeat. Budget your review hours as a line on the bill, because that's where the money goes whichever plan you choose.
Finishing a task doesn't mean you're done with it. Every pull request needs a person to read it, and a wrong one costs you the task, the reading time and the retry.
## Compare Jules with other coding agents
At the start, Jules costs about the same as Claude Pro and ChatGPT Plus, so the sticker price won't separate those three. The daily limits and where the work runs will, and those are what show up on your review hours.
### Price and metering side by side
The table shows each tool's entry plan and the gaps I couldn't fill.
| Tool | Entry price | Top-tier price | Usage metering | Cloud vs local | GitHub integration | Check date |
|---|---|---|---|---|---|---|
| Jules | $19.99/month (Google AI Pro) | not found | not found | Cloud | Yes | not found |
| Codex | $20/month (ChatGPT Plus) | not found | not found | Cloud and local | Yes | not found |
| Claude Code | $20/month (Claude Pro) | not found | not found | Local | not found | not found |
| Copilot agent | $10/month (Copilot Pro) | not found | not found | Cloud | Native | not found |
According to [Plan-Price](https://plan-price.com/en/countries/united-states/), Google AI Pro is $19.99 a month, which is what Jules costs to start.
[Claude](https://claude.com/pricing) Pro and ChatGPT Plus are both $20, so those two sit at the same price, while Copilot Pro costs less.
Half the others, Copilot Pro is $10, but a cheaper plan only helps if its task allowance covers your day, or you're wasting money on insufficient coverage, so the apparent savings vanish if the allowance is too low.
Where the metering column says "not found", I couldn't confirm a figure. Write the number from each vendor's own page on the day you buy, and put that date in the last column.
When a tool runs in the cloud, it does the work on the vendor's machines while you do something else. A local tool runs on your computer and uses your files directly.
Cloud suits tasks you can hand off and review later, but you wait for results and run into daily task caps. Local suits tight back-and-forth, but it ties up your own machine.
### When to choose a Jules alternative
If you live inside GitHub and run few tasks, choose Copilot Pro, because it costs half as much. Choose Claude Code if you want to watch the work happen on your own machine.
When you run many tasks a day, Jules is the wrong pick. In that case its cap, not its price, sets your real cost, so compare the caps before you pay.
## Start Jules tasks automatically with Activepieces
Activepieces has 746 integrations in its MIT-licensed open source core, so a flow that triggers on labelled issues and posts results to Slack is built from ordinary steps. That way you spend your daily task allowance on work you chose rather than on forgotten clicks.
You hit a cap you can't see being used by surprise, and an automation gives you one place to watch it.
### How the Jules task workflow runs
Only one of the four steps spends a Jules task, and everything else is plumbing, leaving you entirely dependent on that single step's efficiency. The diagram that follows shows them in order, with each step handing its result to the next.
1. Activepieces watches your code repository and starts only when someone adds the "jules" label to an issue. Unlabelled issues never use up a task, which is how you keep control of the count.
2. The flow sends the issue title and description to Jules as the instructions. This is the step that counts against your daily cap, so a loosely labelled backlog can drain it fast.
3. A pull request's the proposed code change, waiting for a person to approve it. The flow holds until Jules produces one, so nobody has to keep checking.
4. The team chat gets a Slack message with the link, so the review happens the same day instead of sitting unnoticed.
One caution: whether a ready-made Jules step exists is something you can see by browsing the community integrations folder in the open source repo at github.com/activepieces/activepieces, and how Jules is reached depends on what Google offers on the day you build this, so check Google's own documentation before you commit to a design.
The next part is the one to do first.
Do these three things at the start of the week, in this order, before you spend anything.
1. Open Google's current Jules pricing page and write down the daily task limit and the concurrency limit for each tier, with today's date beside them. Limits change, and a note without a date is a guess.
2. Look at the last week of issues and pull requests, and count how many you would have handed to Jules. Divide by the days you work. If that number sits near a tier's daily cap, you're buying the next tier up whatever the sticker says.
3. Run the labelled-issue flow on the free plan for a few days. If you never hit the cap, you've saved the upgrade. If you do hit it, you now know which limit forced you, and that's the one to pay to lift.
## Frequently asked questions
### Is Google Jules free to use?
Yes, Jules has a free entry level, but it comes with a daily task cap and a limit on how many tasks run at once. If you run a handful of jobs a week, the free level may be all you ever need.
If you run tasks every working day, you'll hit the cap and pay to lift it, so the free price tells you little about your real bill.
### Does Jules have an API, and is it billed separately?
Check Google's pricing page on the day you buy to see whether Jules's programmatic access draws on the same task allowance as your plan.
Until the page says otherwise, assume a task started by a script counts the same as one you start by hand. Assuming the opposite is how a bill surprises you, because an automated job can burn through a day's cap before you've opened your laptop.
Do unused daily tasks roll over? Assume not. Daily allowances are built to reset, not to bank, so a quiet Monday doesn't buy you a heavier Tuesday.
If Google's pages describe any carry-over, that wording overrides this advice.
### Can students or teams get Jules cheaper?
Possibly, but verify it instead of assuming it. Google has attached discounts and promotions to its AI plans for particular groups at various times, and those come and go.
For teams, the safe working assumption is that each person needs their own allowance, so the cost multiplies with headcount. Ask the question on Google's own plan pages before you put a whole group on one tier.
### How often does Google change Jules limits?
Often enough that you should treat any figure as dated, including the ones in this article. **Its caps, tiers and names have moved** as Google has added capacity and features, because Jules is a young product.
The cheap fix is a habit: open Google's pricing and limits pages on the day you buy, write down the date, and look again before each renewal. Skipping that check costs you a plan sized for limits that no longer exist.
## Related reading
- [Zhipu AI API Pricing 2026: Token Costs & Coding Plans](https://www.activepieces.com/blog/zhipu-ai-api-pricing-2026-token-costs-coding-plans)
- [Gemini Agent: What Google Launched, Pricing and Limits](https://www.activepieces.com/blog/gemini-agent-what-google-launched-pricing-and-limits)
- [What Is ZCode? Z.ai's Coding Agent Explained (2026)](https://www.activepieces.com/blog/what-is-zcode-a-developer-guide-to-z-ais-coding-agent)
## References
- [Google Jules](https://jules.google/docs/usage-limits/)
- [Plan-Price](https://plan-price.com/en/countries/united-states/)
- [TestingCatalog](https://www.testingcatalog.com/google-launches-coding-agent-jules-in-beta-with-free-daily-tasks/)