Meta Platforms has officially launched the latest version of its large language model, introducing Llama 4 Scout and Llama 4 Maverick as part of its expanding push into multimodal AI. These new models, described as Meta’s “most advanced yet,” are capable of processing and integrating text, video, images, and audio, pushing the boundaries of content creation and understanding across formats.
Meta emphasized that both Llama 4 Maverick and Scout will be open source, underlining the company’s commitment to accessible innovation. Alongside the release, Meta also previewed Llama 4 Behemoth, calling it “one of the smartest LLMs in the world” and the most powerful model yet, designed to serve as a training benchmark for future AI systems.
Despite the high-profile launch, reports suggest Meta faced setbacks during development. According to The Information, internal benchmarks revealed that Llama 4 lagged behind competitors, particularly in areas such as reasoning, math, and voice interaction—key capabilities where OpenAI’s models still lead.
Nevertheless, Meta is doubling down on AI investments, with plans to allocate up to $65 billion in 2025 toward building out its AI infrastructure. The move comes amid growing investor scrutiny over returns from massive AI spending, as tech giants race to solidify their place in the evolving artificial intelligence landscape.
Source: Swifteradio.com
