跳到正文
SiliconANGLE· Kyt Dotson·· 2 天前AI 評分64

Microsoft 發佈多款自家 AI 模型並更新 Foundry 代理持續改進功能

Microsoft rolls out new first-party models and updates to agentic AI continuous improvement

AI 導讀

Microsoft 推出 Decision-1.1、MAI-Code-2 等自家模型,並更新 Foundry 的代理評估與持續改進功能。Foundry 透過內部評估建立品質、安全、延遲及成本基準,並把專家對代理執行軌跡的評註和優化測試納入改進循環;人工標註目前處於預覽階段。

正文

Microsoft rolls out new first-party models and updates to agentic AI continuous improvement

Microsoft Corp. today rolled out new generative artificial intelligence models and updates to Foundry, the company’s unified enterprise platform that helps developers and organizations ground, test and deploy AI applications and autonomous agents.

Updates to Foundry IQ, part of Microsoft IQ, now provide an enterprise knowledge graph out of the box, delivering a full estate of unstructured and application data to provide rich context for building applications and delivering on essential use cases.

Microsoft also said Foundry now uses existing transparency and observability capabilities to help customers create learning loops. Internal evaluations allow autonomous agents to establish baselines for quality, safety, latency and cost, enabling continuous behavioral improvements.

This includes human annotations, now in preview, that allow experts to rate and comment on agent traces based on business judgment. This feeds back into an improvement loop. This is combined with optimization tests against evaluations and comparisons against real-world results. Foundry now also has new capabilities to adapt to supported models with organizational data to apply reinforcement learning while Foundry runs compute.

Each improvement becomes the next baseline and teams can continue measuring results and test changes, allowing them to improve quality.

Microsoft’s updated decision model and first-party AI models

Now generally available, Microsoft announced Decision-1.1 today, a fast, cost-effective model designed to make reliable decisions in software at scale.

It is a generative AI model that can make decisions dynamically at low cost, rather than relying on hard-coded rules, while offering superior decision-making capabilities compared with large language models. LLMs spend a lot of time making decisions in routing, classification, prioritization, verification, and workflow control by “thinking it through,” which is unnecessary. Decision models are designed to quickly reach answers within constrained definitions and provide calibrated confidence scores, allowing applications to defer when uncertainty is high.

The new model builds on the previous generation, Microsoft-Decision-1, released in October. The new model delivers a significant improvement in intelligence and speed at a competitive price.

The company said it is built on a new base, Google LLC’s Gemma 4 26B A4, adding multimodal support so it can analyze images in instructions.

Microsoft also extended its MAI, a Microsoft AI first-party family of models added to Foundry for enterprise trust.

These models include the new MAI-Code-2, which can handle demanding agentic coding work; MAI-Code-2-Flash supports high-volume, everyday coding, so developers can choose between performance and cost. Both are now generally available on GitHub, with Code-2-Flash in private preview in Foundry.

MAI-Image-2.7 and MAI-Image-2.7-Flash are now in preview and offer sharper editing and more consistent image generation. These will help brands maintain a consistent look across campaigns at a lower cost.

In cybersecurity, MAI-Cyber-2-Flash is a purpose-built model for cyber efforts and available in preview in Codename MDASH – the company’s agentic code security system in preview for eligible organizations.

Microsoft Foundry itself now has an expanded model pool featuring the top 50 models, supporting a model router and AI Gateway for intelligence model selection. The Foundry also offers more than 11,000 models from OpenAI Group PBC, Anthropic PBC, Meta Platforms Inc., Mistral AI SAS, the open-source community and others.

The company said supported models can be evaluated and deployed in Foundry as part of the same continuous improvement workflow used for agents.

Image: SiliconANGLE/Microsoft Designer

A message from John Furrier, co-founder of SiliconANGLE:

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

  • 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more
  • 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network

SiliconANGLE Media is a recognized leader in digital media innovation, uniting breakthrough technology, strategic insights and real-time audience engagement. As the parent company of SiliconANGLE, theCUBE Network, theCUBE Research, CUBE365, theCUBE AI and theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

來源:SiliconANGLE · siliconangle.com