---
title: "oz0 compare start"
description: "Usage and flags for oz0 compare start."
canonical_url: "https://docs.oz0.ai/reference/cli/oz0/compare/start"
---
# oz0 compare start

> Usage and flags for oz0 compare start.

For: Operators and developers looking up a command.

Compare AGENT as installed with one change, on its eval cases (--cases evals, graded by each case's expect) or its recent inputs (--cases recent, passed when the job completes; a data subject's job is never a case). The variant is:

revision=candidate:\<id>       a private candidate of its plugin (its owner or an admin)
revision=gate                 the candidate its plugin's quality gate tests now
revision=rollout              its plugin's rollout's candidate
revision=workspace:\<revision> a saved revision of its plugin's workspace folder
revision=version:\<version>    a version of its plugin the tenant had
draft=\<folder>                your studio draft of its plugin
withhold-skill=\<skill>        the installed agent without one of its skills ([\<plugin>/]\<skill>)
inline-delegate=\<agent>       the installed agent with one of its plugin's delegates folded in: the delegate's
instructions as a skill named after it, its tools beside the agent's, and no
oz.delegate to it (refused when the two hold the sides of a duty pair)

A draft, a workspace revision and a past version run as a private candidate of the plugin on the nodes --nodes names, for the comparison's life, and those nodes take no other work of the tenant meanwhile. Agents on any runtime may be compared; a best-of may not, since its candidates are jobs of their own.

## Usage

```bash
oz0 compare start AGENT --variant revision=<ref>|draft=<folder>|withhold-skill=<skill>|inline-delegate=<agent> --budget USD [flags]
```

## Flags

<table>
<thead>
  <tr>
    <th>
      Flag
    </th>
    
    <th>
      Type
    </th>
    
    <th>
      Default
    </th>
    
    <th>
      Description
    </th>
  </tr>
</thead>

<tbody>
  <tr>
    <td>
      <code>
        --budget
      </code>
    </td>
    
    <td>
      float
    </td>
    
    <td>
      
    </td>
    
    <td>
      what the comparison may spend, in dollars (required)
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --cases
      </code>
    </td>
    
    <td>
      string
    </td>
    
    <td>
      <code>
        evals
      </code>
    </td>
    
    <td>
      evals (the agent's eval cases) or recent (inputs of its jobs that completed in the last 7 days)
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --confidence
      </code>
    </td>
    
    <td>
      float
    </td>
    
    <td>
      <code>
        0.95
      </code>
    </td>
    
    <td>
      the intervals' confidence
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --cost-per-run
      </code>
    </td>
    
    <td>
      float
    </td>
    
    <td>
      
    </td>
    
    <td>
      the estimate's cost of one run, when the agent has no recent jobs to estimate from
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --dry-run
      </code>
    </td>
    
    <td>
      bool
    </td>
    
    <td>
      
    </td>
    
    <td>
      only check the change and show the plan
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --json
      </code>
    </td>
    
    <td>
      bool
    </td>
    
    <td>
      
    </td>
    
    <td>
      print JSON instead of a table
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --margin
      </code>
    </td>
    
    <td>
      float
    </td>
    
    <td>
      <code>
        5
      </code>
    </td>
    
    <td>
      same means not worse than the baseline by more than this many points of pass rate
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --max-cases
      </code>
    </td>
    
    <td>
      int32
    </td>
    
    <td>
      <code>
        10, at most 50
      </code>
    </td>
    
    <td>
      with --cases recent, at most this many inputs
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --nodes
      </code>
    </td>
    
    <td>
      stringToString
    </td>
    
    <td>
      <code>
        []
      </code>
    </td>
    
    <td>
      for a draft, a workspace revision or a past version: the labels of the nodes it runs on (k=v,...)
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --poll
      </code>
    </td>
    
    <td>
      duration
    </td>
    
    <td>
      <code>
        2s
      </code>
    </td>
    
    <td>
      with --wait, how often to look
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --runs
      </code>
    </td>
    
    <td>
      int32
    </td>
    
    <td>
      <code>
        3, at most 10
      </code>
    </td>
    
    <td>
      runs per arm and case
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --tenant
      </code>
    </td>
    
    <td>
      string
    </td>
    
    <td>
      <code>
        default
      </code>
    </td>
    
    <td>
      tenant
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --variant
      </code>
    </td>
    
    <td>
      string
    </td>
    
    <td>
      
    </td>
    
    <td>
      the change: revision=&lt;ref>, draft=&lt;folder>, withhold-skill=&lt;skill> or inline-delegate=&lt;agent> (required)
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --wait
      </code>
    </td>
    
    <td>
      bool
    </td>
    
    <td>
      
    </td>
    
    <td>
      wait for the result and print it
    </td>
  </tr>
</tbody>
</table>

## Inherited flags

<table>
<thead>
  <tr>
    <th>
      Flag
    </th>
    
    <th>
      Type
    </th>
    
    <th>
      Default
    </th>
    
    <th>
      Description
    </th>
  </tr>
</thead>

<tbody>
  <tr>
    <td>
      <code>
        --context
      </code>
    </td>
    
    <td>
      string
    </td>
    
    <td>
      <code>
        $OZ0_CONTEXT or &lt;config dir>/orchestrator-zero/context.json
      </code>
    </td>
    
    <td>
      CLI context file written by server init
    </td>
  </tr>
  
  <tr>
    <td>
      <code>
        --log-level
      </code>
    </td>
    
    <td>
      string
    </td>
    
    <td>
      <code>
        info
      </code>
    </td>
    
    <td>
      log level: debug, info, warn or error
    </td>
  </tr>
</tbody>
</table>

See [all `oz0` commands](https://docs.oz0.ai/reference/cli/oz0) and the [global flags](https://docs.oz0.ai/reference/cli/oz0#global-flags).


## Sitemap

See the full [sitemap](https://docs.oz0.ai/sitemap.md) for all pages.