> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cortex-labs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Understand run outcomes and costs

> Look up run progress, human waits, retries, cancellation and attributed usage.

A logical run records the task. Attempts record individual executions. A retry adds an attempt; it does not erase earlier work or usage.

## Progress and outcomes

| State or outcome | Meaning | Check |
| - | - | - |
| Pending or queued | Admitted work is waiting for execution or capacity. | Dependencies and queue state. |
| Running | One or more attempts are executing. | Current block, input and attempt. |
| Awaiting a person | A question, plan or review needs its permitted decision. | The specific assigned item. |
| Blocked or refused | A required authority, input or limit prevents an operation. | Recorded reason and responsible owner. |
| Failed | An attempt or required path failed. | Retained outputs, failed inputs and retry effects. |
| Skipped | The configured route or input handling excluded work. | Whether it was expected. |
| Partial | Some useful work exists while required coverage remains incomplete. | Missing input and the conclusions it limits. |
| Completed | The configured execution reached its completion conditions. | Review and publication outcomes separately. |
| Cancelling or unresolved | Stop was requested, but execution or effects still need reconciliation. | Remaining work and external acknowledgements. |
| Cancelled | Execution settled after cancellation. | Accepted evidence and effects that already occurred. |

## Human waits are specific

A question requests information. A Verdict requests a configured decision or form. A shared review assesses an exact subject version. Explorer plan acceptance starts Build work.

These decisions have separate authority and records. Typing a message does not settle a pending Explorer plan. Run completion does not establish output approval.

## Retry and cancellation

A retry uses the run's selected versions and records another attempt. Check which outputs already exist before repeating a step.

**A whole model-step retry can repeat tool effects.** Inspect calls that collected data, wrote resources or contacted a destination. Reconcile uncertain external outcomes before starting a replacement action.

Cancellation stops new work and requests child cancellation. It does not remove bytes already delivered externally or promise automatic compensation. Accepted evidence and incurred usage remain.

## Budgets and actual usage

| Measure | Interpretation |
| - | - |
| Inference | Actual model calls, including supporting calls and configured delegated work. |
| Compute | Allocated runtime for agents, scripts and other processing. |
| Storage | Retained source, output and protected evidence under the applicable policy. |
| Estimate | A prediction, not a final charge. |
| Pending price | Measured usage awaiting a price. It is not zero usage. |
| Reused result | No new inference call for the reused computation. Retained storage can still meter. |

The context meter describes the size of the latest model context. It is not the turn's total billed token count.

The root-run budget governs its configured work and child work. Attempts and revisions use remaining allowance. Inspect the configured limit and actual usage rather than assuming a retry has a fresh budget.

An app's model-spend cap is a soft stop between blocks. A running block can exceed it before the next check. It does not by itself cap every storage or compute charge.

A human wait with no allocated compute creates no runtime charge. Retained evidence continues to use storage. Failed work retains incurred usage under the billing policy.

Use [Set budgets and billing ownership](/administration/configuration-and-usage/budgets-billing-ownership) for resource attribution.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.