| status | implemented | ||
|---|---|---|---|
| area | model-generation | ||
| platforms |
|
Prepare the initial language-model session early to reduce perceived startup work when the first prompt is sent.
As a user, I want the first response to begin promptly so that starting a new chat feels responsive.
- The initial model session is prepared when the chat model is created.
- Prewarming does not send a prompt or add a visible chat message.
- Prewarming remains fully on-device and does not require a network request.