Add support in prepareStep for changing the model temperature, maxOutputTokens, topP

[remorses]

I have different values for temperature etc based on the model. I would like to return this data in prepareStep so I can change it based on the current model. This is very important when the model is changed in a step in the middle of a message, otherwise you can't really update the model because parameters like temperature cannot be changed

prepareStep({ model }) {
  if (typeof model === 'object' && model.provider === 'moonshot') {
    return {
      temperature: 0.6,
      topP: 0.48,
    }
  }
  return {
  }
},

[github-actions[bot]]

This discussion was automatically locked because it has not been updated in over 30 days. If you still have questions about this topic, please ask us at community.vercel.com/ai-sdk