惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

美团技术团队
阮一峰的网络日志
阮一峰的网络日志
T
The Blog of Author Tim Ferriss
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
宝玉的分享
宝玉的分享
L
LangChain Blog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
Last Week in AI
Last Week in AI
博客园 - 司徒正美
M
MIT News - Artificial intelligence
人人都是产品经理
人人都是产品经理
WordPress大学
WordPress大学
B
Blog RSS Feed
H
Hackread – Cybersecurity News, Data Breaches, AI and More
博客园 - Franky
B
Blog
V
V2EX
J
Java Code Geeks
D
Docker
博客园 - 叶小钗
The Cloudflare Blog
量子位
博客园_首页
MongoDB | Blog
MongoDB | Blog

Recent Announcements

Amazon EC2 X8i instances are now available in Europe (Paris) Amazon CloudWatch pipelines now supports drop and conditional processing AWS Deadline Cloud supports monitor creation in multiple regions Amazon CloudWatch pipelines introduces new compliance and governance capabilities Second-generation Amazon FSx for NetApp ONTAP is now available in four additional AWS commercial and AWS GovCloud (US) Regions AWS Billing and Cost Management Dashboards Now Supports Scheduled Email Delivery AWS RTB Fabric supports health checks for real-time bidding workloads AWS Backup extends Amazon FSx support to 5 additional AWS Regions and expands cross-Region and cross-account copy to 14 AWS Regions Amazon RDS now supports the latest CU and GDR updates for Microsoft SQL Server Amazon Timestream for InfluxDB Now Supports Customer-Defined Maintenance Windows Amazon Bedrock now supports cost allocation by IAM user and role Amazon OpenSearch Service supports Managed Prometheus and agent tracing Amazon S3 Lifecycle pauses actions on objects that are unable to replicate Amazon RDS Blue/Green Deployments now supports Amazon RDS Proxy AWS Marketplace announces the Discovery API for programmatic access to catalog data AWS Agent Registry for centralized agent discovery and governance is now available in Preview Amazon OpenSearch Serverless now supports Zstandard (zstd) codec for index compression AWS Private CA now supports customer managed permissions for cross-account sharing Amazon EC2 Capacity Manager now supports tag-based dimensions Amazon Route 53 Resolver endpoints now support DNS delegation for private hosted zones in AWS GovCloud (US) Regions SageMaker HyperPod now supports gang scheduling for distributed training workloads Amazon IVS 实时直播功能现已支持冗余摄取 Amazon EKS 托管节点组现在支持 EC2 Auto Scaling 暖池 Amazon Bedrock AgentCore 浏览器新增操作系统级交互功能 Amazon WorkSpaces Advisor now available for AI-powered troubleshooting Amazon OpenSearch Service now supports Graviton4 based i8ge instances Oracle Database@AWS is now available in twelve AWS Regions AWS Lambda expands response streaming support to all commercial AWS Regions AWS Cost Explorer launches Natural Language Query capabilities powered by Amazon Q Amazon Lightsail 现已在亚太地区(马来西亚)区域推出
Amazon SageMaker JumpStart 现推出两种用于代理式编码和高效...
Amazon Web Services · 2026-05-15 · via Recent Announcements

今日,AWS 宣布在 Amazon SageMaker JumpStart 中推出 GLM-5.1-FP8 和 Phi-4-mini-instruct,进一步扩展了面向 AWS 客户的基础模型产品组合。这些模型由 Z.ai 和 Microsoft 开发,为在 AWS 基础设施上运行的企业人工智能工作负载带来了先进的代功能与高效的推理性能。

这些模型利用专业功能应对不同的企业人工智能挑战:

GLM-5.1-FP8 在代理式软件工程领域表现卓越,具备持续的多轮优化能力,可处理存储库级别的代码生成、终端任务以及复杂的调试流程,并通过扩展推理能力不断提升性能。它非常适合自动化代码审查管道、人工智能驱动的开发环境,以及需要模型经过数百轮迭代以优化解决方案的长期问题解决场景。

Phi-4-mini-instruct 在内存受限和延迟敏感的环境中擅长强推理、数学和逻辑,支持 24 种语言和函数调用,且采用紧凑型设计。它非常适合边缘部署、对延迟敏感的应用程序、多语言聊天机器人,以及客户需要在最小资源开销下实现强大推理能力的场景。

借助 SageMaker JumpStart,客户只需点击几下即可部署这些模型中的任何一个,以解决其特定的人工智能使用案例。

要开始使用这些模型,请导航到 SageMaker Studio 的“模型”部分,或使用 SageMaker Python SDK 将模型部署到您的 AWS 账户。有关在 SageMaker JumpStart 中部署和使用基础模型的更多信息,请参阅 Amazon SageMaker JumpStart 文档。