The news: Meta launched compressed versions of its Llama AI models, allowing them to run efficiently on smartphones and tablets and leaping over Google and Apple in mobile AI performance.

Tests on OnePlus 12 Android phones revealed that Llama 3.2 1B and 3B models have nearly the same performance as larger versions while being 56% smaller and using 41% less memory.

Pure play AI vs. integrated mobile strategies: Meta’s small models have cracked a key challenge in mobile AI performance without relying on specialized hardware or extensive cloud resources, per VentureBeat.