Microsoft has shifted its focus from frontier AI development to cost-effective models, releasing two new in-house coding models: MAI Code 2 and Code 2 Flash. These models offer competitive performance on coding benchmarks at a significantly lower cost compared to leading models like GPT-6.1 Sol and Opus 5.5. MAI Code 2, a large Mixture-of-Experts model with a 1M context window, achieves high performance for a fraction of the cost, while Code 2 Flash, a smaller distilled model, is even more economical and outperforms comparable models. Both models are integrated into Microsoft's Copilot, with Code 2 Flash available on the free tier. AI
IMPACT Microsoft's move signals a potential industry shift towards more accessible and affordable AI models, impacting enterprise adoption and competition.
RANK_REASON This is a significant product release from a major tech company, focusing on cost-effective AI models rather than pure frontier capabilities. [lever_c_demoted from significant: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →