Choose a model

<table> <thead> <tr> <th>Model</th> <th>Use when</th> <th>Prompting focus</th> </tr> </thead> <tbody> <tr> <td style={{ whiteSpace: "nowrap" }}> [`gpt-realtime-2`](https://developers.openai.com/api/docs/models/gpt-realtime-2)

  <td>
    You need the strongest realtime reasoning, tool use, and instruction
    following.
  </td>
  <td>
    Tune reasoning effort, preambles, tool policies, exact entity capture,
    and long-session state.
  </td>
</tr>
<tr>
  &lt;td style=&#123;&#123; whiteSpace: "nowrap" &#125;&#125;>
    [`gpt-realtime-1.5`](https://developers.openai.com/api/docs/models/gpt-realtime-1.5)
  
  <td>You need a fast, reliable non-reasoning speech-to-speech model.</td>
  <td>
    Follow the core realtime prompt structure and test for latency-sensitive
    behavior.
  </td>
</tr>

</tbody> </table>