TSTickerStarGet alerts

Microsoft announces seven MAI models and MAI Licensed Weights

Microsoft announced a lineup of seven models for cyber defense, coding, image editing, speech, transcription, and model routing, alongside MAI Licensed Weights and the Maia 200 accelerator. Disclosed benchmarks and pricing include MAI Code 2 delivering 98% of GPT-6.1 Sol's coding quality at about 40% of the price and MAI Realtime 1 starting at $0.01 per audio minute. The article does not disclose company financial results, financial guidance, or transaction terms.

Share this story

Key takeaways

  • The seven-model lineup consists of MAI Cyber 2 Flash, MAI Code 2, MAI Code 2 Flash, MAI Image 2.7, MAI Realtime 1, Dragon Copilot Transcribe, and Decision 1.1.

  • Across four coding benchmarks, MAI Code 2 delivered 98% of GPT-6.1 Sol's coding quality at about 40% of the price and 96% of Claude Opus 5.5's performance at 5% of the price.

  • MAI Cyber 2 Flash output tokens are priced 86% below MAI Cyber 1.1 Flash, while MAI models account for almost 40% of MDASH scans.

  • MAI Image 2.7 ranked first on Artificial Analysis image editing, and MAI Image 2.7 Flash is priced at approximately $10 per 1,000 images.

  • Dragon Copilot Transcribe handles 10 million patient encounters per month, powers 100% of doctor-patient conversation transcriptions, and halves multilingual error rates.

  • Decision 1.1 is used by 4,000 customers, while Maia 200 is in production and achieves 1.7 times more tokens per watt.

  • MAI Licensed Weights lets customers customize frontier MAI models with proprietary data and run them in their own Azure environments.

Key metrics

MAI Cyber 2 Flash benchmark price
Less than 2% of GPT-6 Astra's cost per task on the CyberGym-E2E benchmark
MAI Cyber 2 Flash output-token price change
86% lower than MAI Cyber 1.1 Flash
MAI share of MDASH scans
Almost 40%
MAI scan continuation rate in MDASH
90% of scans that started with MAI continued to use it
MAI Code 2 performance and price versus GPT-6.1 Sol
98% of coding quality at about 40% of the price
MAI Code 2 performance and price versus Claude Opus 5.5
96% of performance at 5% of the price
MAI Code 2 Flash comparative pricing
GPT-6 Luna at 65% of the price and Claude Haiku 5.5 at 20% of the price
MAI Image 2.7 Artificial Analysis ranking
#1 for image editing and #4 for text-to-image generation

Primary source

TickerStar generated this brief from an official company, investor-relations, earnings-call, or regulatory source. Review the primary material before making an investment decision.

Open Investor relations

Follow MSFT in real time

Get material earnings, filing, IR, and earnings-call alerts by email.

Create a free alert