Mountain View: Google has reportedly postponed the launch of its next-generation artificial intelligence model, Gemini 3.5 Pro, after the model failed to meet internal performance benchmarks, particularly in coding. The reported delay comes as competitors OpenAI and Anthropic continue to roll out increasingly capable AI models, putting additional pressure on Google in the rapidly evolving AI landscape.
According to a Bloomberg report, Gemini 3.5 Pro was originally expected to launch in June, shortly after Google CEO Sundar Pichai unveiled the model during the company’s Google I/O developer conference in May. At the event, Pichai had said that Google had already begun using Gemini 3.5 Pro internally.
However, sources familiar with the development told Bloomberg that engineers have been dissatisfied with the model’s progress, especially in coding capabilities, leading to a delay in its public release.
Coding performance reportedly falls short
The report states that Google’s engineering teams have grown increasingly frustrated with Gemini 3.5 Pro’s inability to reach the company’s internal expectations.
Coding has become one of the most competitive areas in generative AI, with enterprises increasingly relying on AI-powered software development tools. OpenAI has expanded its coding platform Codex, while Anthropic has continued to strengthen its position in enterprise AI through its Claude family of models.
According to Bloomberg, Google updated Gemini’s training data late last month in an attempt to improve its coding performance. Despite the effort, the improvements reportedly failed to satisfy internal evaluation standards.
Employees cited in the report expressed concern that Google risks losing further ground to rivals if Gemini continues to lag behind in one of AI’s fastest-growing use cases.
Competition in AI continues to intensify
The reported delay comes at a time when competitors have accelerated the release of more advanced AI models.
OpenAI has introduced GPT-5.6, while Anthropic has launched Claude Fable 5, both of which have attracted significant attention for their capabilities across reasoning, coding and enterprise applications.
According to the report, the sophistication of these models was such that US authorities initially requested both companies to limit certain aspects of their releases while safety assessments were carried out.
The rapid pace of innovation from competitors has heightened expectations for Google’s next flagship model and increased scrutiny over any delays.
Alphabet shares decline
Following Bloomberg’s report, investors reacted negatively to the news.
Shares of Alphabet, Google’s parent company, fell nearly 3 per cent, with losses reaching around 4 per cent during Thursday’s trading session before recovering slightly.
The market reaction reflected investor concerns about Google’s competitive position in artificial intelligence, an area that has become central to the future growth strategies of major technology companies.
Internal challenges slow development
Beyond technical performance, the report suggests Google’s organisational structure has also contributed to development delays.
Employees said multiple business units are involved in AI model launches as Google seeks to integrate generative AI across products including Google Search, Maps, YouTube, Android and Google Cloud.
One former employee described coordinating these efforts as being like trying to “boil an ocean”, highlighting the complexity of aligning numerous teams under a common strategy.
Current and former staff also cited shifting priorities, duplicated work across departments and competition for computing resources as factors slowing product development.
These internal challenges have reportedly made it more difficult for Google to bring new AI products to market quickly.
Google says testing is ongoing
Responding to Bloomberg’s report, a Google spokesperson said the company continues to develop and test multiple AI models.
“We’re shipping quickly across a wide range of models while keeping them highly cost-effective for customers,” the spokesperson said.
The company added that it is currently testing Gemini 3.5 Pro, an upgraded Flash model and several other AI systems with partners.
Google also confirmed that it remains engaged with the US government as regulatory scrutiny surrounding advanced AI models continues to increase.
Technology companies developing frontier AI systems are facing growing expectations around safety testing, national security assessments and responsible deployment before releasing new models publicly.
Google retains strengths in multimodal AI
Despite the reported delay, analysts and researchers believe Google continues to possess important strengths in artificial intelligence.
Gemini benefits from Google’s extensive search infrastructure, allowing it to retrieve and integrate information from the company’s vast search ecosystem.
Google has also highlighted its progress in multimodal AI, with capabilities spanning text, images and video, alongside research into so-called “world models” designed to simulate real-world environments.
The company has invested heavily in integrating AI across its products since the launch of ChatGPT in late 2022 prompted Google to declare a “code red” to accelerate AI development.
Google co-founder Sergey Brin has reportedly played an active role in encouraging faster progress in AI coding technologies.
AI strategy continues to evolve
According to the report, Google initially restricted employees from using Gemini to generate or analyse software code due to concerns that proprietary code might inadvertently enter future training datasets.
Those restrictions have since been relaxed.
Google recently disclosed that approximately 75 per cent of the code written within the company is now generated using AI before being reviewed by engineers prior to deployment.
The company has also consolidated several coding initiatives under its Google Antigravity project, although employees reportedly continue to face limitations due to competition for computing capacity.
Meanwhile, competition for AI talent has intensified, with some Google researchers reportedly leaving to join Anthropic and other AI laboratories.
Conclusion
The reported delay of Gemini 3.5 Pro underscores the increasingly competitive nature of the global AI race. While Google remains a leading player with strengths in search integration and multimodal AI, challenges in improving coding performance and coordinating development across multiple teams have reportedly slowed the launch of its flagship model. As rivals continue to release more advanced AI systems, the timing and capabilities of Gemini 3.5 Pro will be closely watched by both the technology industry and investors.
