Google delays Gemini 3.5 Pro launch as AI model struggles to meet expectations: Report
Follow UsGoogle has reportedly delayed the launch of its next major AI model, Gemini 3.5 Pro, because it is not yet meeting the company's internal expectations.
The delay has likely created frustration inside the company.
Some employees are reportedly worried that rivals like OpenAI and Anthropic are moving ahead more quickly in the AI race.
Google has reportedly delayed the launch of its next major AI model, Gemini 3.5 Pro, because it is not yet meeting the company’s internal expectations. According to a report by Bloomberg, the model is now running months behind schedule as Google continues to improve its performance before releasing it. The report, which cites people familiar with the matter as well as ten current and former Google employees, says the delay has created frustration inside the company. Engineers, researchers, and managers are reportedly disappointed by the slow progress. Some employees are also worried that rivals like OpenAI and Anthropic are moving ahead more quickly in the AI race.
SurveyGemini 3.5 Pro was widely expected to be announced during Google’s developer conference in May. However, the company reportedly decided not to launch the model because it was not ready.
Also read: OpenAI is selling ChatGPT Basketball, but why?
One of the biggest challenges appears to be coding performance. According to the report, Google updated the model’s training data last month to improve its coding abilities. However, the changes reportedly did not produce the results the company had hoped for.
Bloomberg also reports that Google’s large size may be making development more difficult. Different teams across Google Cloud, DeepMind, Android, and other parts of the company are all working on AI coding tools at the same time. The report says multiple layers of decision-makers are involved before a model can be released, which may slow down the process. Employees are also said to compete for computing resources when using AI systems internally.
Also read: Sam Altman admits OpenAI’s past year was not its best, says it is his fault
However, Google disagreed with the idea that it is falling behind. A company spokesperson said Google is “shipping quickly across a wide range of models” while keeping them cost-effective. The company also confirmed that it is currently testing Gemini 3.5 Pro, an upgraded Flash model, and several other AI models with selected partners. Also, Google said that it is discussing model testing and safety standards with the US government.
Also read: OpenAI enters hardware market with Codex Micro: What is it and how much it costs
