Google Bets on Gemini 4 to Lead the AI Race

Google has confirmed that its next-generation AI model, Gemini 4, will play a major role in the company’s efforts to remain competitive in the rapidly evolving artificial intelligence market.

During Alphabet’s second-quarter 2026 earnings call, CEO Sundar Pichai acknowledged that while Google continues to perform strongly across many areas of AI, improvements are still needed in coding and agentic coding. He explained that the company believes a larger and more capable Gemini 4 model will be essential for competing at the highest level of AI development.

Coding Remains a Key Focus

Pichai admitted that coding is one of the areas where Google still has work to do.

Although the company has made significant progress with its Gemini models, it sees advanced coding capabilities and AI agents that can independently complete complex tasks as important priorities for future development.

Google recently released Gemini 3.6 Flash, which includes improvements to coding performance and efficiency. However, Pichai suggested that meaningful progress at the next stage of AI innovation will require the more advanced architecture currently being developed for Gemini 4.

Earlier this year, he also acknowledged that Google had fallen slightly behind competitors in agentic coding, partly because rival companies have gained valuable user feedback through developer-focused AI tools already available on the market.

Gemini 3.5 Pro Still Delayed

While Google continues to release updates to its Flash models, its flagship Gemini 3.5 Pro model has yet to launch.

The company originally introduced the Gemini 3.5 family during Google I/O in May, releasing Gemini 3.5 Flash immediately while indicating that the Pro version would follow shortly afterwards.

That release has since been delayed.

Google says Gemini 3.5 Pro is currently being tested with selected partners and will be released once it meets the company’s quality standards. Reports suggest that coding performance has been one of the main reasons behind the delay, with recent improvements failing to reach Google’s internal expectations.

As a result, there is still no confirmed public release date for the model.

New Flash Models Arrive

Although the flagship model remains unavailable, Google has continued expanding its AI portfolio.

The newly launched Gemini 3.6 Flash is designed to become Google’s everyday workhorse model. According to the company, it produces around 17% fewer output tokens than Gemini 3.5 Flash while improving coding performance and overall efficiency.

Google has also introduced Gemini 3.5 Flash-Lite, a faster and more affordable option aimed at handling high-volume AI workloads. The model is already being integrated into Google Search, helping deliver quicker and more efficient AI-powered responses.

Gemini 4 Is Already in Development

Alongside the latest product launches, Google confirmed that Gemini 4 has entered its pretraining phase.

Pretraining is one of the earliest stages of building a large language model, during which the system learns from vast amounts of data before moving on to more specialised training and fine-tuning.

Google describes Gemini 4 as one of its most ambitious AI projects to date. However, the company has not announced when the model will be available to developers or the public.

Competition in AI Continues to Intensify

The AI race continues to accelerate as major technology companies invest heavily in larger, more capable models.

Google has faced increasing pressure to keep pace with competitors, particularly in coding-focused AI tools. Recent reports have also highlighted the departure of several senior AI researchers from DeepMind, adding further attention to Google’s long-term AI strategy.

Despite these challenges, the company continues to release regular updates while preparing its next major generation of models.

What This Means for Users

For businesses and developers already using Google’s AI products, Gemini 3.6 Flash provides immediate improvements in speed, efficiency and coding performance.

However, much of the industry’s attention is now focused on the delayed release of Gemini 3.5 Pro and the future arrival of Gemini 4.

Whether Google can deliver these models on schedule will be closely watched, as successful launches will demonstrate its ability to remain competitive in one of the fastest-moving areas of technology.

Looking Ahead

Google’s latest announcements show that the company is taking a long-term approach to AI development.

While Gemini 3.6 Flash offers meaningful improvements today, Google’s leadership has made it clear that Gemini 4 represents the next major milestone in its AI roadmap. Until then, developers and businesses will be watching closely for news on the release of Gemini 3.5 Pro and further updates on Gemini 4’s progress.