Google Delays Gemini 3.5 Pro AI Model Amid Coding Performance Concerns

By: Harsh Vardhan

|

Updated on: 17-Jul-2026 04:00 PM

7,902 views

Follow Us:

7,902 views

​Google has delayed the launch of its next flagship AI model, Gemini 3.5 Pro, after the model failed to meet internal performance goals, especially in coding. This setback comes as competitors OpenAI and Anthropic continue to advance their own AI models, maintaining a lead in coding capabilities.

Key Highlights

• Google delayed Gemini 3.5 Pro after the model failed to meet internal coding performance goals.
• Competitors OpenAI and Anthropic have released advanced AI models, increasing pressure on Google.
• Alphabet shares fell nearly 3 percent after news of the delay and internal frustrations surfaced.
• Google says 75 percent of its code is now generated by AI and reviewed before production.
• Internal complexity and resource competition have contributed to slower progress on Gemini 3.5 Pro.

Gemini 3.5 Pro Faces Development Challenges

According to a Bloomberg report, Gemini 3.5 Pro was originally scheduled for release in June. Google CEO Sundar Pichai had previously announced the model and stated that internal testing had already begun. However, Google engineers have expressed dissatisfaction with the model’s performance, particularly when compared to rival offerings.

OpenAI and Anthropic have released new AI models that have drawn attention from U.S. authorities due to their advanced capabilities. In contrast, Google’s Gemini 3.5 Pro has struggled to match their progress, especially in coding tasks. One source said that Google updated Gemini’s training data in an attempt to improve coding performance, but the results did not meet expectations.

The importance of coding in AI models has grown, with OpenAI focusing on its Codex platform to target the enterprise market, an area where Anthropic also has a strong presence. Google’s delays have raised concerns among employees about the company’s ability to compete in this critical area.

Internal Frustrations and Organizational Hurdles

Alphabet, Google’s parent company, saw its shares fall by nearly 3 percent following news of the delay, with trading dropping as much as 4 percent at one point on Thursday. The report highlights growing frustration among Google engineers, AI researchers, and managers over the slow progress of Gemini 3.5 Pro.

Employees cited organizational complexity as a major factor in the delays. Multiple teams, including Google Cloud, Google DeepMind, and Android, are involved in AI development, making coordination challenging. One former employee compared the process to trying to “boil an ocean.” Shifting priorities, duplicated efforts, and competition for resources have made it harder to maintain a clear strategy and bring products to market quickly.

Google has also faced early restrictions on using Gemini for coding tasks due to concerns about proprietary code leaking into training data. Although these policies have been relaxed, former staff said they limited early experimentation and slowed progress.

Current Status and Future Prospects

A Google spokesperson told Bloomberg that the company is testing Gemini 3.5 Pro and other models with partners. The spokesperson emphasized that Google aims to ship models quickly while keeping them cost-effective. Google is also working closely with U.S. authorities as scrutiny of advanced AI models increases.

Despite the setbacks, Google continues to highlight Gemini’s strengths, such as its ability to query Google search data and its work across text, images, and video. The company has also streamlined some coding tools under the Google Antigravity project. Currently, 75 percent of code at Google is generated by AI and reviewed before deployment.

Internal frustration over Google’s AI position has led to some departures to competitors like Anthropic. Access to Anthropic’s Claude model within Google is now limited to select research and high-priority teams.