The AI model to use for error explanation.

<p><strong>Recommended models by provider:</strong></p>
<ul>
<li><strong>OpenAI</strong>: 
  <ul>
    <li>gpt-5-nano (default, cost-effective)</li>
    <li>More advanced models available https://platform.openai.com/docs/models/</li>
  </ul>
</li>
<li><strong>Google Gemini</strong>:
  <ul>
    <li>gemini-2.0-flash (default, fast and efficient)</li>
    <li>More advanced models available https://ai.google.dev/gemini-api/docs/models</li>
  </ul>
</li>
<li><strong>Ollama</strong>:
  <ul>
    <li>gemma3:1b (default, fast and efficient)</li>
    <li>More advanced models available https://ollama.com/search</li>
  </ul>
</li>
</ul>

<p>The model you choose affects both the quality of explanations and the API cost. Generally, more advanced models provide better explanations but cost more per request.</p>