Skip to main content
Traders' Hub
Back to blog
AI Technology

Microsoft Introduced New AI Models

ნუცა ტყეშელაშვილიNovember 18, 20261 min read
Microsoft Introduced New AI Models

Microsoft has unveiled a new suite of artificial intelligence models featuring improved capabilities in coding, cybersecurity, and other areas. With this move, the company is seeking to compete with Anthropic and OpenAI.

At the Microsoft Ignite conference held in San Francisco, the company introduced a new coding model, MAI Code 2. According to Microsoft, its performance nearly matches that of competing AI models, though it comes with significantly lower operating costs.

The company also presented a model specialized in cybersecurity, which will initially be available to a limited group of users. Its purpose is to help detect software vulnerabilities and reduce the risks associated with them. Microsoft estimates that as AI tools continue to advance, security flaws in software systems are also being exposed more frequently.

In addition, the company released new AI models for image generation and audio-to-text transcription.

Microsoft's strategy and security priorities

Mustafa Suleyman, head of Microsoft AI, told Reuters that building specialized technologies tailored to different fields allows the company to achieve results comparable to — and in some cases better than — its competitors, at a lower cost.

"This isn't just one model, and it's not a one-off success. We've built a system that continuously improves results while simultaneously optimizing for both performance and price," Suleyman said.

He noted that recent cybersecurity incidents, including a case in which OpenAI's AI agents breached the Hugging Face platform, have shown that focusing solely on performance when developing models is not enough — special attention must also be paid to safety mechanisms.

According to Suleyman, his top priority is building mechanisms that prevent artificial intelligence from engaging in harmful actions and ensure that models behave according to their intended rules. This area of AI research is known as alignment, and it refers to ensuring that AI behavior is consistent with human goals and safety requirements.

Bulls & Bears in your inbox

Get our latest market analysis and financial insights delivered straight to you. No spam — just the signal.