Microsoft's AI division, MAI, has developed seven new in-house models for various applications including coding, security, and language processing. These models are designed to be more cost-effective than external alternatives, with MAI Code 2 achieving nearly the same performance as GPT-6.1 Sol but at a significantly lower cost. The new models are intended for continuous operation within enterprise environments, particularly for AI agents. AI
IMPACT Microsoft's development of cost-effective in-house models could reduce reliance on external providers and accelerate AI agent deployment in enterprises.
RANK_REASON Frontier lab (Microsoft) announces new in-house models with performance and cost comparisons. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →