Microsoft's new MAI Code 1.1 Flash gets crushed by Deepseek on both price and performance

Microsoft has introduced MAI Code 1.1 Flash, a specialized model for GitHub Copilot designed to increase token efficiency and lower operational costs by 75 percent compared to its predecessor. Despite these gains, performance benchmarks indicate that the model remains less cost-effective and slower than Deepseek V4 Flash. This release highlights the intensifying competition in the code-generation market, where proprietary offerings are increasingly evaluated against lower-cost alternatives from other developers.