Native multimodality, mixture-of-experts models, super long context windows, step changes in performance, and unparalleled efficiency. All in easy-to-deploy sizes custom fit for how you want to use it. We've optimized models for easy deployment, cost efficiency, and performance that scales to billions of users. We can’t wait to see what you build.

2025年4月5日 · The Llama 4 model collection also supports the ability to leverage the outputs of its models to improve other models including synthetic data generation and distillation. The Llama 4 Community License allows for these use cases. Out-of-scope: Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use ...

LLaMa 沿着小 LLM 配大数据训练的指导思想，训练了一系列性能强悍的语言模型，参数量从 7B 到 65B。 例如，LLaMA-13B 比 GPT-3 小10倍，但是在大多数基准测试中都优于 GPT-3。 大一点的 65B 的 LLaMa 模型也和 Chinchilla 或者 PaLM-540B 的性能相当。 同时，LLaMa 模型只使用了公开数据集，开源之后可以复现。 但是大多数现有的模型都依赖于不公开或未记录的数据完成训练。 1.3 LLaMa 预训练数据. LLaMa 预训练数据大约包含 1.4T tokens，对于绝大部分的训练数 …

2024年12月11日 · The Large Concept Model is trained to perform autoregressive sentence prediction in an embedding space. We explore multiple approaches, namely MSE regression, variants of diffusion-based generation, and models operating in a quantized SONAR space.

Llama 2 is now accessible to individuals, creators, researchers, and businesses of all sizes so that they can experiment, innovate, and scale their ideas responsibly. This release includes model weights and starting code for pre-trained and fine-tuned Llama language models — ranging from 7B to 70B parameters.

2024年4月26日 · First things first—what is Llama 3? It is a text-generation open-source AI model that takes in a text input and generates a relevant textual response. It is trained on a massive dataset (15 trillion tokens of data to be exact), promising improved performance and better contextual understanding.

2024年12月10日 · 美东时间12月6日，Meta 在 X 平台宣布推出最新开源大型 语言模型 Llama-3.3-70B。 Llama 3.3 采用优化的 transformer 架构，融合了监督式微调（SFT）和基于人类反馈的 强化学习 （RLHF）等先进技术。 支持 128K tokens的上下文长度，约等于 400 页文本。 在多个行业基准测试中， Llama-3.3-70B 的表现超过了谷歌的 Gemini 1.5 Pro、OpenAI 的 GPT-4o 和亚马逊的 Nova Pro，展现出了强大的竞争力。 虽然 Llama 3.3 只有 700亿 参数，但在 性能 上已 …

2024年5月24日 · In this article, we explore the revolutionary advancements of the LLaMA series, from its inception to the cutting-edge capabilities of LLaMA 3, and its transformative impact on AI language...

2024年4月24日 · In this article we will discuss the core concepts behind Llama 3, explore its innovative architecture and training process, and provide practical guidance on how to access, use, and deploy this groundbreaking model responsibly.

2025年4月6日 · According to Meta, “Llama 4 Scout is best-in-class on image grounding, able to align user prompts with relevant visual concepts and anchor model responses to regions in the image.”