Meta has officially released its Llama 4 AI Multimodal Models. The models power the Meta AI assistant on the web. Additionally, it also powers the WhatsApp, Messenger, and Instagram. You can download the model ether from Meta or Hugging Face.
The official blog post reads, “Llama 4 Maverick, a 17 billion active parameter model with 128 experts, is the best multimodal model in its class, beating GPT-4o and Gemini 2.0 Flash across a broad range of widely reported benchmarks, while achieving comparable results to the new DeepSeek v3 on reasoning and coding—at less than half the active parameters. Llama 4 Maverick offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena.”
The Llama 4 AI is a small model that is capable of fitting in the single Nvidia H100 GPU. Amazon has also announced the availability of Llama 4 AI models on its platform. The blog post says, there are two models including Llama 4 Scout 17B and Llama 4 Maverick 17B and they are able to understand both image and text prompts.
Amazon blog post says, “The availability of Llama 4 Scout and Llama 4 Maverick on AWS expand the already broad selection of models offered to customers to build, deploy, and scale their applications. AWS consistently offers new models from leading AI companies such as Meta as soon as the models are released, with enterprise-grade tools and security that make it easy to build, customize, and scale generative AI applications.”
The Llama 4 Scout 17B comes with a 17 billion active parameter model with 16 experts. It can fits on a single H100 GPU.However, the Llama 4 Maverick comes with 17 billion active parameter model with 128 experts and it can fits on a single H100 host. The company also trained a teacher model known as Llama 4 Behemoth. The company says it has outperformed GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on STEM-focused benchmarks such as MATH-500 and GPQA Diamond.
Get latest Tech and Auto news from Techlusive on our WhatsApp Channel, Facebook, X (Twitter), Instagram and YouTube.
