---
title: "Model Status"
slug: "determining-a-models-status"
updated: 2025-01-09T15:53:19Z
published: 2025-01-09T15:53:19Z
---

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

# Model Status

### When should you trust a model's prediction?

#### Model Status

The model's status value is critical for determining if the model is in a safe reliable state. This status value can be [mapped to the control system](/v1/docs/mapping) using a system tag and then used to determine if the predicted value should be used. The below chart gives the definition for each model state:

| Value | State | Meaning |
| --- | --- | --- |
| 0 | Inactive | The model is disabled, this either means the model has not been enabled in Koios, or the Predict Engine service is not running. |
| 1 | Running | The model is enabled and has no errors. The prediction can be used as needed. |
| 2 | Failed | The model is enabled but has an error associated with it. This could be an error on a binding, a tag, or the model in general. Please check the error code for more detail. |

### Using this in logic

Any time a prediction is used in the control system, you should also take into account that the model is in a Running state. This is commonly done with some form of an "Equal" block. If the model is in a failed or inactive state, it is common practice to do the following:

- Trigger an alarm to the HMI
- Fallback to traditional control

![](https://cdn.document360.io/9fab5c8f-2756-48da-b963-232bd8533649/Images/Documentation/Model%20Status.png)