Verified65% confidenceFactTime unknown
Meta推出的Llama 3.1拥有从8B到405B的多个参数版本，覆盖从边缘部署到数据中心的全场景需求
3
Sources
65%
Confidence
Medium-term (~90 days)
Relevance
6/1/2026
First Seen
Valid until: 8/30/2026
Sources
企业大模型选型指南：Llama3.1、Qwen2.5、DeepSeek深度对比
bilibili云栖智联
Related Entities
Related Claims
UnverifiedMeta发布的Llama 3.1已支持405B参数规模，并以宽松的商业授权开放79% similarUnverifiedMeta的Llama 3.2系列推出了1B和3B的轻量版本73% similarUnverifiedMeta的LLaMA 3系列最大规模为405B，DeepSeek-V3约为671B（混合专家架构）71% similarUnverifiedLlama.cpp was originally created with the goal of running Meta's LLaMA model on pure CPU environments67% similarUnverifiedMeta发布了名为Muse Glimmer 30B的本地大语言模型，采用Apache 2.0许可证，拥有296亿参数，是一个稠密模型64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/18064API
curl https://kongchang.com/api/v1/knowledge/claims/18064MCP
get_claim(id=18064)