Wednesday, 18 November 2026
Rīga TV

World and Latvian news in one place

TechnologyPublished: 18 November 2026 at 22:04

Microsoft's MAI unit releases seven new AI models at lower prices

Microsoft's AI division MAI unveiled seven new models at Ignite 2026 covering coding, cybersecurity, images, speech and decision routing. The company stresses lower prices compared with OpenAI and Anthropic models.

Foto: Hacker News (≥150 p.)

Microsoft's artificial intelligence unit MAI announced seven new AI models on November 17 at its Ignite 2026 conference. They cover coding, cybersecurity, image processing, speech and conversation, and decision routing. The company says price is the main selling point: as spending on tokens (the units in which models are billed) rises, customers want models that let them run agents around the clock.

Coding

According to Microsoft, MAI Code 2 reaches 98% of the quality of OpenAI's GPT-6.1 Sol across four benchmarks at about 40% of the price. Against Anthropic's Claude Opus 5.5, it reaches 96% at 5% of the price. The company acknowledges the model is not yet tuned for harder tasks such as fundamental science research and very long-horizon agentic work. It is available in GitHub Copilot, while the lighter MAI Code 2 Flash is also in limited preview on Microsoft Foundry.

Cybersecurity, images and voice

MAI Cyber 2 Flash, the company says, beats GPT-6 Astra's quality on the CyberGym-E2E benchmark at less than 2% of the cost per task. MAI Cyber accounts for almost 40% of scans in MDASH.

MAI Image 2.7 ranks first in image editing on Artificial Analysis and fourth in text-to-image generation. MAI Realtime 1 is the company's first live speech-to-speech model, with pricing starting at $0.01 per audio minute. Dragon Copilot Transcribe, built on MAI Transcribe 2, handles 10 million doctor–patient encounters a month and has halved error rates in multilingual workflows.

Other announcements

The Decision 1.1 model is, according to Microsoft, 50 times faster and 79 times cheaper than GPT-6.1 Sol. The company also said its second-generation AI accelerator, Maia 200, is now in production and delivers 1.7 times more tokens per watt. A new option, MAI Licensed Weights, will let customers customize an MAI model with their own data and run it in their own Azure environment.

All comparisons are figures provided by the company.

Comments

0/1500

Comments are automatically moderated. No hate, threats, personal data or spam.

Loading comments…

More in this category