WhatsApp for Android is tipped to be working on a new feature that will let users choose the Large Language Model Meta AI (Llama) 3 model to power the Meta AI chatbot. On Thursday, a feature tracker claimed that this feature is currently under development and not available to beta testers. The feature is said to let users choose between the Llama 3-70B AI model and the Llama 3-405B AI model. The latter is tipped to be available for a limited number of prompts.
WhatsApp may introduce option to choose AI Model
According to WhatsApp feature tracker WABetaInfo, the new feature was spotted in the WhatsApp for Android beta version 2.24.14.7. However, it is not visible at present and beta testers will not be able to interact with it. This feature builds on top of the recent introduction of the Meta AI chatbot in the app and lets users choose from the smaller Llama 3-70B model and the latest and larger Llama 3-405B model.
WhatsApp’s Meta Llama Model feature
Photo Credit: WABetaInfo
In a screenshot shared by the feature tracker, a new settings menu can be seen titled Meta Llama model. This space will let users choose the AI model. The above mentioned options are shown with the Llama 3-70B model listed as the “faster” model and the Llama 3-405B model listed as “Better for more complex prompts”.
A block of text underneath states, “Preview the 405B model for a limited number of prompts each week. Once you’ve reached the limit, you can continue your conversations using the 70B model.”
Notably, the Llama 3-405B AI model has not yet been released to the public. While releasing the 8B and 70B Llama 3 models in April, Meta stated that it is still training the family’s largest model with 400 billion parameters. If the information about the new WhatsApp feature is true, it is likely that the company intends to fine-tune the model based on user interactions.
As stated in a previous report, Meta has already confirmed that it is training its AI models with user data during interactions with its AI chatbots.
Affiliate links may be automatically generated – see our ethics statement for details.
