A pay model in which each accepted unit of work — one annotation batch, one test, one transcription file — pays a fixed amount. It is the dominant model in microtask and AI-training work. Per-task pay is transparent (the price is shown before you start) but unforgiving: rejected work typically pays nothing, and your effective hourly rate collapses whenever the task queue runs dry.
Last reviewed: September 2026 · Part of the PaidTaskers glossary
How per-task pay works
Per-task pay means each accepted unit of work — one annotation batch, one usability test, one transcription file — pays a fixed amount that is shown before you start. It is the dominant pay model in microtask and AI-training work, and it sits behind most entry-level online task platforms.
The word doing the heavy lifting is 'accepted'. On most platforms, payment depends on the work passing review, so the sticker price of a task is a ceiling, not a guarantee. Acceptance is what converts a task price into money, which is why this model rewards precision as much as speed.
Per-task pay versus hourly pay
Per-task pay is transparent but unforgiving; hourly pay is steadier but usually lower. With per-task work you always know the price of the next unit, yet your income stops completely whenever tasks run out, and a rejected task typically pays nothing at all.
Hourly arrangements — common in moderation and search-evaluation contracts — pay for scheduled time regardless of queue depth, which makes them easier to plan around. The trade-off is that hourly work usually requires an application, a qualification test, and a commitment to shifts, while per-task work is open to almost anyone immediately.
How per-task pay affects your real earnings
Your real earnings under per-task pay are the task price multiplied by your acceptance rate, divided by all time worked including dead time. Rejected work typically pays nothing, so a platform advertising $2 batches becomes a very different proposition at a 70 percent acceptance rate.
The second drag is queue droughts. Your effective hourly rate collapses whenever the task queue runs dry, because time spent hunting for available work is unpaid. That is why the same platform can feel like decent money on a stocked day and a wasted afternoon when work is thin, with no change in your effort or skill.
How to earn more from per-task work
The reliable levers are acceptance rate and dead time, not raw speed. Reading guidelines before starting, checking your first batch against feedback, and working when queues are stocked all raise the share of work that survives review — the unpaid rejection page covers the defenses in detail.
Speed matters too, but only after accuracy: a fast contributor who fails quality benchmarks loses project access, which is the most expensive outcome of all. Treat your first tasks on any new platform as paid practice, and expect your rate per hour to improve as the guidelines become familiar.
How PaidTaskers presents per-task pay
Platform pages in the AI training hub label per-task pay explicitly and pair it with what is known about acceptance standards and queue availability. Reviews in this directory keep the advertised task price separate from the realistic hourly outcome, because conflating the two is the most common way beginners overestimate this work.
The microtask entry covers the entry-level end of this model, where prices are cents and volume is everything.
Frequently asked questions
Is per-task pay better than hourly pay?
Neither is universally better: per-task pay offers immediate access and transparent prices, while hourly pay offers steadier income and protected time. Per-task work favors fast, accurate workers on platforms with deep queues; hourly work favors anyone who wants predictable earnings. Compare both on effective hourly rate, not on the sticker price.
Do I get paid if my task is rejected?
Usually not — rejected work typically pays nothing on per-task platforms, which is the model's sharpest edge. Platforms differ in how much feedback they give and whether partial credit exists at all. Reading guidelines carefully and matching your work against provided examples is the practical defense against unpaid rejections.
Why do tasks keep disappearing on per-task platforms?
Because task availability swings with client demand, your region, and how many other contributors are online at the same moment. Queue droughts are a normal feature of the model, not a personal penalty. Experienced contributors spread across two or three platforms so a dry queue on one is not a lost day.
What is a realistic hourly rate from per-task work?
Entry-level microtasks often land at a few dollars per hour, while accurate contributors on better-paying per-task platforms can do materially better once skilled. The honest method is to track accepted earnings divided by all time spent — including searching and rejections — over at least a week before drawing any conclusions.
Used by ai training & data labeling platforms
Browse the ai training & data labeling hub