Meta reports that Muse Spark achieves its reasoning capabilities using over an order of magnitude less compute than Llama 4 Maverick, its previous mid-size flagship.

The new model is competitive with rival AI models, according to data from Meta, but won’t be widely available outside Meta’s own product ecosystem.

Meta on Tuesday announced the latest version of its Llama AI model, Llama 3.1. The newest Llama technology comes in three different versions, with one variant being the biggest and most capable AI model from Meta to date. Like previous versions of Llama ...

Meta’s latest open source AI model is its biggest yet. Today, Meta said it is releasing Llama 3.1 405B, a model containing 405 billion parameters. Parameters roughly correspond to a model’s problem-solving skills, and models with more parameters ...

A legal case involving Meta revealed the company's secret experiments with training data. Meta used "ablation" to identify how specific data improved its Llama AI models. Some researchers say this could support a system to assign value to AI data and pay ...