Technical details
What's the model? We use OpenAI's ChatGPT 3.5-turbo-16k as our default model to generate responses. It's suitable for most basic use cases, such as FAQ AIs. If you need an AI with more advanced reasoning capabilities, GPT-4 is a better option. However, we currently can't offer GPT-4 for free because it's approximately 20x more expensive than GPT-3.5. If you're interested in using GPT-4 and can cover the associated costs, please contact us at hi@glome.social
Can I use other models like Claude?
At the moment, we use only OpenAI models. In the near future, we might integrate other services and models. The features we prioritize largely depend on user demand: the more requests we get for a feature, the higher it ranks on our to-do list.
Context window
Our context window is set at 16k, which equates to about 12k English words. The size of this context window is reduced by: - the soul prompt - any AI description you've provided - any knowledge base you've provided - all conversation messages in the current conversation
Last updated