Key facts
- Microsoft unveiled MAI Code 2, a programming model it says nearly matches rivals at a fraction of the cost.
- Microsoft plans to offer a cyber-specific model to a limited group of customers.
- Microsoft AI chief Mustafa Suleyman said domain-specific models help keep costs low while matching rivals.
- Suleyman cited OpenAI agents' hacking of the Hugging Face platform as shaping his focus on safeguards.
Microsoft announced a suite of AI models with coding and cybersecurity capabilities on November 17, 2026, at its Ignite conference in San Francisco, saying it is racing to challenge rivals including Anthropic and OpenAI.
The centerpiece is a computer programming model called MAI Code 2, which Microsoft says almost matches the performance of rival AI models at a fraction of the cost. The company also touted a cyber-specific model it plans to offer to a limited group of customers, aimed at helping address a rise in software vulnerabilities that AI tools are increasingly bringing to light. Microsoft released new image-generation and transcription AI models as well.
Mustafa Suleyman, Microsoft's AI chief, told Reuters in an interview that building technology specific to different domains was helping the company keep costs low while matching rivals and in some cases topping model benchmarks. "This isn't just one model. It's not a flash in the pan," he said. "This is a system, a hill-climbing machine that we have built, which co-optimizes performance and price."
Suleyman said the recent cybersecurity reckoning, illustrated by OpenAI agents' hacking of the Hugging Face platform, had led him to focus not just on performance but on critical safeguards. Preventing AI from taking harmful actions and ensuring it behaves as intended, a research area known as alignment, was his top priority, he said.