Google Launches Gemini 3.6 Flash, Delays Gemini 3.5 Pro, Begins Gemini 4 Pretraining

The CODEW | AI Watch

Google has introduced three new additions to its Gemini family—Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber—while confirming that its flagship Gemini 3.5 Pro remains delayed after missing multiple development milestones. At the same time, the company revealed it has already begun pretraining Gemini 4, signaling that its next-generation frontier model has become Google's primary strategic priority.

The announcement reflects a notable shift in Google's AI strategy. Rather than competing solely on benchmark performance, the company is emphasizing lower inference costs, production efficiency, enterprise deployment, and broad ecosystem adoption while preparing Gemini 4 for future competition against OpenAI, Anthropic, and other frontier AI developers.

Gemini 3.6 Flash Focuses on Cost Efficiency

Among the new releases, Gemini 3.6 Flash is positioned as Google's primary production model for developers and enterprises.

Google reduced pricing to $1.50 per million input tokens and $7.50 per million output tokens, compared with the previous Flash generation's $9 per million output tokens. The company also reported that Gemini 3.6 Flash uses approximately 17% fewer output tokens on the Artificial Analysis Index, potentially lowering inference costs for organizations operating high-volume AI workloads.

The model's knowledge cutoff has also been updated from January 2025 to March 2026, giving developers access to more recent information without requiring retrieval-based augmentation for many applications.

Security-Focused Gemini Flash Cyber

Google also introduced Gemini 3.5 Flash Cyber, a specialized version designed for cybersecurity operations.

Unlike Google's general-purpose AI models, Flash Cyber will be available only to governments and selected trusted partners. The restricted deployment reflects growing industry efforts to balance advanced defensive AI capabilities with concerns over dual-use technologies that could be exploited for offensive cyber operations.

The launch positions Google more directly against specialized enterprise security offerings from competitors such as Anthropic and Microsoft as AI-assisted cyber defense becomes an increasingly strategic market.

Google Looks Beyond Gemini 3.5

Perhaps the most significant announcement was Google's confirmation that pretraining has already begun for Gemini 4.

Although Gemini 3.5 Pro remains under development, Google's decision to publicly discuss its next-generation model suggests the company is already shifting attention toward a broader platform refresh rather than extending the lifecycle of the current generation.

The strategy indicates that Google is prioritizing long-term platform competitiveness while continuing to improve deployment economics for existing enterprise customers.

Enterprise Adoption Continues to Accelerate

Despite increasing competition across the AI market, Google's distribution ecosystem continues expanding rapidly.

According to company figures:

• The Gemini application surpassed 900 million monthly active users, more than doubling from the previous year.
• AI Overviews now reach more than 2 billion monthly users across over 200 countries and territories.
• Google reports more than 8 million paid Gemini Enterprise seats across 2,800 organizations, with paid usage increasing approximately 40% quarter over quarter.

Google's enterprise momentum has also been strengthened through strategic partnerships, including its reported agreement with Apple to provide a custom Gemini model for the next generation of Siri expected to launch with iOS 27.

Industry Perspective

Google's latest product announcements suggest the company is repositioning its AI strategy around practical deployment rather than solely pursuing benchmark leadership.

Lower pricing, improved token efficiency, and broader enterprise distribution indicate that Google is targeting production workloads where operational costs and scalability increasingly influence customer purchasing decisions. At the same time, early investment in Gemini 4 demonstrates that competition among frontier AI developers continues to accelerate, with future model capabilities expected to remain a key differentiator.

For enterprise customers, the latest Gemini releases provide additional options optimized for cost, security, and large-scale deployment while offering an early indication of Google's longer-term roadmap for its next generation of AI systems.

Editorial Sources

Primary Sources

• Google — Introducing Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber

Additional Reporting & Analysis

• Build Fast With AI — AI News Today (July 22, 2026): Gemini 3.6 Flash, Gemini 4 Pretraining, and Enterprise AI Updates.
• Stan Ventures — State of Google Gemini (July 2026).
• Business Insider — Google Doubles Down on Cheaper, Faster AI While Delaying Gemini 3.5 Pro.

The CODEW verifies information using official announcements and multiple independent, authoritative sources before publication whenever possible.

AI Watch is The CODEW's recurring intelligence report covering the companies, technologies, investments, infrastructure, policies, and market trends driving enterprise artificial intelligence. Rather than reporting individual announcements, AI Watch explains what those developments mean for businesses, investors, and the broader technology industry.