Chatbot API Sandbox Page

Keep in mind that this page is mostly intended for testing purposes.

Loading info...

[Click RUN to generate text]

  Generally, output takes 15-30 seconds for the GPT-2 models and up to a minute for the GPT-Neo model, depending on how much it is warmed up.
  It is possible for outputs to be invalid, restarting the model and timing out. However, if three consecutive attempts with basic inputs fail, you can safely assume that the model is currently not working.