Google has delayed the release of Gemini 3.5 Pro, its next-generation flagship AI model, as the company works to improve its performance, particularly in coding tasks, according to a Bloomberg report citing people familiar with the matter. Alphabet shares fell about 4% on Thursday following the report.
Google CEO Sundar Pichai had announced at the company's I/O developer conference in May that Gemini 3.5 Pro would launch in June. That timeline has now slipped, with the company continuing to refine the model before release.
According to Bloomberg, the delay has led to frustration among some engineers and AI researchers, who are concerned about Google's pace as competitors continue to introduce new AI models.
The report says Google is working to improve Gemini 3.5 Pro's coding capabilities before launch. Coding has become one of the most competitive areas in generative AI, with models from companies including OpenAI and Anthropic widely used for software development.
Bloomberg also reported that Google is consolidating multiple internal AI coding initiatives into a more unified system, a process that may be contributing to the delay.
A Google spokesperson said the company is currently testing Gemini 3.5 Pro, an updated Gemini Flash model, and other AI systems with selected partners. The company also said it continues to work with the US government on AI testing and evaluation.
Google highlighted its broader AI progress, including advances in multimodal AI systems capable of understanding text, images and video, as well as research into AI world models.
The reported delay comes as competition in generative AI intensifies, with OpenAI, Anthropic and Meta continuing to advance their AI models. The market reaction highlights how closely investors are tracking AI product launches as they become increasingly important to major technology companies.