Google delays Gemini 3.5 Pro launch as coding performance falls short of internal goals

5 Sources

Share

Google's most powerful AI model, Gemini 3.5 Pro, is now months behind schedule as engineers struggle to improve its coding capabilities. The delay has sent Alphabet shares down 4% and raised concerns about Google losing ground to rivals like OpenAI and Anthropic in one of AI's most critical commercial battlegrounds.

Google Faces Setback as Gemini 3.5 Pro Launch Slips by Months

Google is grappling with a significant delay in releasing Gemini 3.5 Pro, its most powerful flagship AI model, as the tech giant works to address performance shortfalls that have frustrated internal teams and spooked investors

1

. The AI model delay has pushed the launch several months behind schedule, with underwhelming coding capabilities at the heart of the problem

2

. The setback comes at a critical juncture when competition from OpenAI and Anthropic has intensified, with both companies recently releasing models that outperform Google's current offerings in generating software code

3

.

Alphabet shares fall 4% following the Bloomberg report, highlighting growing investor skepticism about Google's ability to maintain its edge in the rapidly evolving AI landscape

2

. The stock dropped to $354.17, reflecting market concerns about whether the tech giant can hold onto market share against increasingly aggressive rivals

3

.

Source: Benzinga

Source: Benzinga

Organizational Complexity Hampers Google's AI Development Challenges

The delay in Gemini 3.5 Pro launch stems from deeper structural issues within Alphabet. Multiple teams across Google—including DeepMind, Cloud, Android, and Search—are simultaneously developing AI coding tools, leading to overlapping efforts, shifting priorities, and slower execution

3

. This organizational complexity, combined with multiple layers of stakeholders involved in preparing models for release and integrating AI across a vast product portfolio including search, maps, and YouTube, has created bottlenecks that competitors appear to avoid

4

.

Recent attempts to improve coding performance through updated training data have reportedly fallen short of expectations. Late last month, Google updated the data being used to train Gemini to enhance these skills, but the results were disappointing according to sources familiar with the matter

4

. The setback has frustrated Google engineers, AI researchers, and managers, with many concerned the company risks losing its competitive position

4

.

Competition Heats Up as Rivals Advance in Code Generation

The pressure on Google intensifies as competitors continue rolling out more advanced models. OpenAI recently launched GPT-5.6 Sol AI, with CEO Sam Altman claiming the model improves token-processing efficiency by 54% in programming tasks carried out by AI agents

5

. Meta unveiled Muse Spark 1.1 last week, described by its chief AI officer as the group's best-performing model for coding tasks and AI agents

5

. Anthropic and several Chinese labs including Z.ai are also intensifying competition with open-weight models available through the open-source ecosystem

5

.

Google's flagship AI model was originally announced in May at the Google I/O developer conference, with the company stating it was being used internally but wouldn't be ready for broader rollout until the following month

2

. That timeline has now slipped considerably, raising questions about Google's execution capabilities in enterprise AI applications.

Google Maintains Testing Progress Despite Setbacks

Despite the challenges, a Google spokesperson defended the company's pace, telling CNBC that the company is "shipping quickly across a wide range of models while keeping them highly cost-effective for customers"

2

. The spokesperson added that Google is currently testing 3.5 Pro, an upgraded Flash model, and other models with partners, and is productively engaged with the US government on model testing and broader frameworks.

The engagement with the US government reflects increased regulatory scrutiny of AI companies' most advanced models, with authorities monitoring capabilities and industry safety standards

4

. This adds another layer of complexity to Google's already challenging rollout process, as the company must balance performance improvements with regulatory compliance and safety considerations. The delay underscores the pressure bearing down on leading artificial intelligence players as labs attempt to deliver increasingly capable models for software development use cases while keeping operating costs under control

5

.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved