Llama - Maginative

Meta Llama

Meta Splits AI Division to Speed Up Product Rollouts Amid Talent Exodus

Chris McKay• May 26, 2025 • 3 min read

Meta is restructuring its AI division into two teams to accelerate development amid fierce competition, but faces a massive talent exodus with 11 of 14 original Llama researchers having left the company.

Meta Llama

Meta Launches Llama 4: Multimodal, Massive, and Made for Everyone

Chris McKay• April 5, 2025 • 3 min read

Meta has released Llama 4 Scout and Maverick, two open-weight AI models designed for multimodal reasoning, with Maverick outperforming GPT-4o and Scout offering a record-breaking 10M token context window.

Meta Llama

Meta’s Llama AI Model Hits 1 Billion Downloads

Chris McKay• March 18, 2025 • 2 min read

Meta’s Llama AI model has surpassed 1 billion downloads, reflecting its widespread adoption across research, development, and commercial applications, despite ongoing challenges in regulation, licensing, and competition.

Meta Llama Open Source

Meta Wraps Up 2024 with the Release of Llama 3.3

Chris McKay• December 6, 2024 • 2 min read

This new model is more efficient, cost-effective, and accessible for developers seeking high-quality text-based AI solutions.

Meta Llama

Meta Releases Llama 3.2: Lightweight and Multimodal Models

Chris McKay• September 25, 2024 • 3 min read

Two smaller models, 1B and 3B parameters, are designed for edge and mobile devices, while larger 11B and 90B parameter models bring vision capabilities to the Llama ecosystem.