Gemini3

Gemini 3.5 Pro has not yet appeared, Google may skip waiting and turn to Gemini 4. On August 11, the latest report from industry analysis agency SemiAnalysis stated that Google may have shelved the originally planned Gemini 3.5 Pro model and may fill the market gap caused by the delay of the flagship model by promoting the Gemini 4 series models in advance. Previously, Google has attracted attention due to the delayed release of Gemini 3.5 Pro. Outside reports indicate that the model still needs to be optimized in some key capabilities, especially code-related performance. At the Google I/O developer conference in May 2026, Alphabet CEO Sundar Pichai said that Gemini 3.5 Pro is expected to be launched in June, but the model has not yet been officially launched. Later in Alphabet's second quarter earnings conference call, Pichai continued to emphasize Google's progress in the field of cutting-edge AI and expressed confidence in the development of the Gemini series models. SemiAnalysis pointed out that the current Gemini series has faced competitive pressure in some AI capability evaluations, and believes that its development speed lags behind some cutting-edge models. Previously, it was reported that Gemini3.5Pro briefly appeared on the Arena platform and was subsequently removed about 30 minutes after it went online, triggering discussions about the status of model testing. According to reports, Google’s previous delay in Gemini 3.5 Pro was mainly related to the optimization of model capabilities, including the performance of enterprise-level application scenarios such as code generation. Google has said that the model is still in the testing phase and is being evaluated with partners. If SemiAnalysis's judgment is true, Google may be adjusting the Gemini product rhythm and investing more resources in the next-generation Gemini4 series to cope with the continued accelerating model iteration speed of OpenAI, Anthropic and other AI laboratories. Currently, competition in large models has shifted from simply releasing new versions to a comprehensive competition of capabilities, costs, application ecology, and R&D efficiency. via AI News (author: AI Base)