Tech giant Meta has made a significant move in the artificial intelligence sector by introducing the latest member of its Llama family, Llama 4. Set to be released in 2025, this model aims to automate complex multi-step tasks and emulate human-like problem-solving abilities. This development aligns with Meta's ambition to lead the AI race.
The Llama 4 series starts with two models: Llama 4 Scout and Llama 4 Maverick, designed to assist users in creative tasks such as general assistance and chat. According to Meta, Maverick excels in areas like software coding and multilingual reasoning, outperforming competing models such as OpenAI’s GPT-4 and Google’s Gemini 2.0 in specific benchmarks. CEO Mark Zuckerberg announced that two more models, Llama 4 Behemoth and Llama 4 Reasoning, will be unveiled shortly.
As a multimodal AI system, Llama 4 can interact with both text and images. The Scout model features a 10-million-token context window, enabling it to perform complex tasks like multi-document summarization and processing user activity for personalized tasks. Meta indicates that Llama 4's performance claims position it favorably against advanced open-source models.
These advancements illustrate Meta's commitment to enhancing its competitive edge in AI technology and enriching user experiences.