Meta is gearing up to launch Llama 4, its next flagship AI model, which will emphasize enhanced voice features. Expected in the near future, this model aims to allow users to interrupt its speech output, similar to functionalities in competitors' offerings. As voiced by Meta's chief product officer, Llama 4 will be an all-encompassing model capable of handling speech, text, and various data types. The pressure from competing Chinese AI lab DeepSeek, which has demonstrated superior performance, has prompted Meta to accelerate its development efforts.
Meta chief product officer Chris Cox stated Llama 4 will be an 'omni' model, capable of natively interpreting and outputting speech as well as text and other types of data.
According to reports, Meta is working to enhance voice capabilities with Llama 4, allowing users to interrupt the model mid-speech, similar to other innovations in the field.
The competitive landscape is heating up as open models from the Chinese AI lab DeepSeek are performing on par or better than Meta's Llama models, prompting accelerated development.
Meta is setting up war rooms to understand how DeepSeek reduced costs in model deployment, indicating a strategic pivot in their approach to AI model development.
Collection
[
|
...
]