독서리서치레이더투자 프레임워크
로그인 / 회원가입
← 읽기로 돌아가기
뉴스주제 · Microsoft MAI Model Family중대2 분 소요DisconfirmAI 편집부

Microsoft Unveils Seven In-House AI Models and Weight Licensing Service at Ignite 2026, Challenging OpenAI and Anthropic on Cost

Microsoft launched seven MAI models covering code, security, image, and voice, plus a weight licensing service for private Azure deployment.

2026-11-18 10:34 GMT+8

US tech giant Microsoft announced on November 17, 2026, at its Ignite conference that it has developed and released seven in-house AI models through its Microsoft AI (MAI) division. These models cover programming, cybersecurity, image generation, real-time voice, medical transcription, and model routing. Mustafa Suleyman, CEO of Microsoft AI, stated that as enterprise token spending rises rapidly, Microsoft aims to provide solutions with performance approaching top-tier models from OpenAI and Anthropic but at significantly lower costs.

In core programming tasks, Microsoft introduced MAI Code 2 and MAI Code 2 Flash. According to Microsoft, MAI Code 2 achieves 98% of the performance of OpenAI's GPT-6.1 Sol across four benchmarks while costing only 40% as much per task; compared to Anthropic's Claude Opus 5.5, it reaches 96% performance at just 5% of the cost. MAI Code 2 Flash targets high-volume repetitive work, with costs approximately 65% of GPT-6 Luna and 20% of Claude Haiku 5.5.

Beyond coding, the new portfolio includes MAI Cyber 2 Flash for continuous monitoring and vulnerability scanning, which Microsoft claims performs comparably to GPT-6 Astra in CyberGymE2E tests at less than 2% of the cost. For images, MAI Image 2.7 ranks first on the Artificial Analysis leaderboard, priced at one-fifth of competitor GPT Image 2.6 Sunburst. Microsoft also released MAI Realtime 1 for low-latency dialogue and Dragon Copilot Transcribe, a medical transcription service using the MAI-Transcribe-2 model, which is already used by about one-third of physicians in the US.

Notably, Microsoft also launched Decision 1.1, a model routing system that automatically selects the most appropriate AI model based on task requirements to avoid sending all workloads to expensive large models. Microsoft states that this system has been adopted by 4,000 customers, runs 50 times faster than GPT-Sol 6.1, and costs approximately 1/79th as much.

로그인 / 회원가입
독서리서치레이더투자 프레임워크

In addition to API services, Microsoft announced the "MAI Licensed Weights" service. This initiative allows enterprises to acquire trained model weights and deploy them within their own Azure environments, customizing them with internal data. This means enterprises do not need to rely entirely on Microsoft-managed APIs, giving them control over access permissions, proprietary data, and customization outcomes.

출처

  • ithome.com.twiThome↗
주제

Microsoft MAI Model Family

Microsoft has launched a new strategic phase with its in-house AI model family, MAI (Microsoft AI), marking a shift from…

3 단계1 건업데이트됨 2026-11-20 22:36

현재까지의 과정

  1. 2026-06-02June 2, 2026: Microsoft AI announces the initial MAI model family at Build 2026, including MAI Thinking 1, MAI Code 1 Flash, and MAI Image 2.5, establishing the foundation for in-house development.microsoft.ai ↗
  2. 2026-11-17November 17, 2026: At Ignite 2026, Microsoft releases seven updated/new MAI models (Code 2, Cyber 2 Flash, Image 2.7, Realtime 1, etc.) and introduces the MAI Licensed Weights service for private Azure deployment.ithome.com.tw ↗
  3. 2026-11-18November 18, 2026: Industry analysts note that the pricing strategy shifts competition from raw benchmark scores to unit-task cost, with MAI Code 2 claiming 98% of GPT-6.1 Sol's performance at 40% of the cost.stock.finance.sina.com.cn ↗

아직 해결되지 않은 문제

  • Will enterprises adopt MAI Licensed Weights over managed APIs, and how will this impact Microsoft's cloud revenue mix?
  • Can MAI models maintain competitive parity with future OpenAI/Anthropic releases without distillation, given the rapid pace of frontier research?
  • How will the integration of Maia 200 accelerators affect the actual latency and throughput benefits claimed for MAI Realtime and Code models?
전체 주제 읽기 →
리서치 계속하기 →