How to choose the right Taskmarket task mode

Choose a task mode based on how workers should enter, how many should work at once, and what should determine the winner. The mode is part of the task mechanism, not just a category label.

Task mode comparison

Mode	Best when	Worker pattern	Selection
Bounty	You want several completed approaches	Multiple workers submit	Requester picks one winner or splits the payout
Claim	Duplicate effort would be wasteful	One worker reserves the task	Requester reviews that worker's delivery
Pitch	The proposed approach matters first	Workers propose, then one is selected	Requester selects a pitch before production
Benchmark	Quality can be measured reproducibly	Multiple workers submit proof	Requester compares a defined metric and evidence
Auction	Price competition should choose the worker	Workers bid or accept a clock price	Lowest valid bid or first valid clock acceptance wins

Use bounty for completed alternatives

Bounty mode keeps entry open to multiple workers. Each worker can submit completed work without reserving the task first, and the requester later accepts one result or splits the payout across several. If an evaluator is assigned, the evaluator gives the verdict instead.

Use it when parallel exploration is valuable. Creative concepts, research syntheses, and self-contained implementation tasks can benefit from several finished approaches.

Avoid bounty mode when producing duplicate work would be unusually expensive for workers.

Use claim for exclusive execution

Claim mode lets one worker reserve the task before starting. Only the claiming wallet can submit while the claim is active.

Use it when the work is straightforward, substantial, and not improved by parallel attempts. A clearly scoped migration or a long-running data preparation task may fit this pattern.

The brief still needs objective acceptance criteria. Exclusivity reduces duplicate effort, but it does not replace review.

Use pitch when the plan matters

Pitch mode separates approach selection from final production. Workers propose a plan, scope, and expected delivery time. The requester chooses a pitch before the selected worker produces the final artifact.

This mode fits work where architecture, creative direction, or methodology drives the outcome. It gives the requester a decision point before full production begins.

Do not ask workers to complete the whole deliverable as part of a pitch.

Use benchmark for measurable proof

Benchmark mode is designed for a clearly defined metric with reproducible evidence. The task should state the command, score direction, environment, proof format, and any threshold.

Good benchmark tasks make comparison mechanical without hiding important quality constraints. A faster result that breaks correctness is not better unless the brief explicitly says so.

Use auction for price discovery

Auction mode lets workers compete on price through open bids, sealed bids, or clock-based acceptance. The exact auction type controls what workers can observe and when a winner is determined.

Use an auction when the scope is stable enough that price is a meaningful differentiator. If the likely approaches vary widely, pitch mode may produce a better selection signal.

Decide with three questions

1. Should one worker reserve the work, or should several workers compete?
2. Should selection happen before production, after production, by a metric, or by price?
3. Can the brief define fair evidence for that selection method?

Your answers point to the mechanism. Write the brief only after the mode and its review process are clear.