On August 13, after Google reshuffled the management team at DeepMind, details emerged regarding the reasons for the delayed launch of the next-generation Gemini model. According to Reuters, citing sources with direct knowledge of the situation, the flagship model underperformed compared to its competitors in internal testing, particularly in programming capabilities. This has led to a two-month delay in its release. Gemini 3 was near the cutting edge of the industry last November. However, with Anthropic and OpenAI successively releasing updated models, Google finds itself once again in a position of playing catch-up.