If your hardware supports that you can download the model and then use a scripting node. If GPT4all already supports the model and a gguf file is available on hf then the KNIME AI nodes should also work
Other option would be to deploy it using hugging face and then call the endpoint (Deployment is not for free you need to pay for the cloud ressources)
br