Meta releases Llama 4 AI; Claims edge over GPT-4o, Claude ...
Meta Platforms announced the launch of its latest large language models, Llama 4 Scout and Llama 4 Maverick, solidifying its position in artificial intelligence. Llama is a cutting-edge multimodal AI system capable of handling diverse data formats including text, images, video, and audio. This versatility allows for content conversion across various formats, setting it apart from conventional AI models.
Advancements in AI Capabilities
Llama 4 Scout and Llama 4 Maverick have been hailed by Meta as the best multimodal models globally in their class. What sets them apart is their open-source availability, allowing developers worldwide to leverage the technology for innovative applications. Additionally, Llama 4 Behemoth, positioned as one of the most intelligent language models globally, surpasses benchmarks set by competitors like GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro in various STEM fields.
Performance and Accessibility
Llama 4 Scout, equipped with 17 billion active parameters and 16 experts, outshines its predecessors while operating efficiently on a single NVIDIA H100 GPU. On the other hand, Llama 4 Maverick, with the same number of parameters but 128 experts, exhibits superior performance over GPT-4o and Gemini 2.0 Flash in public benchmarks. The model matches DeepSeek v3 in reasoning and coding tasks using less than half the number of active parameters.
Future Investments
Llama 4 Comparison with Claude 3.7 Sonnet, GPT-4.5, and Gemini 2.5 ...
Meta's commitment to AI development is evident in its plans to invest up to $65 billion this year to expand its AI infrastructure. This strategic move aligns with the growing demand for demonstrable returns on AI investments from major tech companies.
Stay updated with the latest news and developments by subscribing to news alerts.
