Microsoft has introduced MAI Code 1.1 Flash, an updated code model designed for GitHub Copilot. According to The Decoder, this new model improves token efficiency by 25 percent and reduces costs to a quarter of its predecessor's price.
Despite these improvements, The Decoder reports that MAI Code 1.1 Flash faces strong competition from Deepseek's V4 Flash model, which outperforms Microsoft’s offering in both price and performance benchmarks.
This development highlights the rapidly evolving AI coding assistant market, a trend that Japanese tech and investment sectors are closely monitoring as demand for cost-effective and efficient coding tools grows in the region.