Google DeepMind has unveiled three new additions to its Gemini family: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber. The new releases are designed to deliver faster performance, lower operating costs, and greater reliability for developers and enterprises building AI agents at scale.
The announcement highlights Google’s continued focus on production-ready AI models optimized for coding, multimodal workloads, and cybersecurity while intensifying competition with OpenAI and Anthropic.
Gemini 3.6 Flash Delivers Better Performance at Lower Cost
Google describes Gemini 3.6 Flash as its new workhorse model, offering significant improvements in coding, knowledge work, and multimodal understanding.
According to the company, the model reduces token consumption by up to 17% compared to Gemini 3.5 Flash, allowing developers to lower inference costs while maintaining fast response times and strong performance.
Gemini 3.6 Flash is designed for production applications such as AI assistants, coding tools, enterprise automation, and intelligent agents that require high throughput and cost efficiency.
Gemini 3.5 Flash-Lite Targets Cost-Sensitive Workloads
Google also introduced Gemini 3.5 Flash-Lite, its most cost-efficient Flash model to date.
The model is intended for high-volume, lightweight AI tasks, including text summarization, information extraction, classification, and other routine workloads where speed and affordability are essential.
Gemini 3.5 Flash Cyber Focuses on Cybersecurity
The company also launched Gemini 3.5 Flash Cyber, a specialized model fine-tuned to identify and help remediate cybersecurity vulnerabilities.
Google said the model will be available exclusively to governments and trusted partners through a limited-access pilot program, reflecting its growing investment in AI-powered cybersecurity solutions.
Built for AI Agents
Google DeepMind said the primary goal of these releases is to improve efficiency, reduce latency, and increase reliability for customers deploying AI agents in production environments.
The new Flash models are designed to support organizations running large-scale AI applications that require fast responses, consistent performance, and lower operating costs.
Gemini Pro Update Still Missing
Despite launching three new models, Google did not announce the long-awaited update to Gemini Pro, its flagship model for advanced reasoning and coding tasks.
Gemini Pro was last updated in February, although Google previously indicated that a new version was being tested internally. Recent reports have suggested that the company delayed the release while working to meet internal performance targets.
Competition Continues to Intensify
The launch comes as competition in the AI industry continues to accelerate.
In recent months, OpenAI released GPT-5.5 and began rolling out GPT-5.6, while Anthropic introduced Claude Opus 4.8 and Claude Sonnet 5 and expanded access to its Fable 5 frontier model. The rapid pace of innovation has increased pressure on Google to advance its flagship AI offerings.
Gemini 3.5 Pro and Gemini 4 Are on the Horizon
Google DeepMind Product Lead Logan Kilpatrick confirmed that Gemini 3.5 Pro is currently being tested with partners and is expected to launch soon.
He also revealed that the company has begun its most ambitious pre-training run yet for Gemini 4, signaling Google’s continued investment in developing more capable next-generation AI models.
Read the article in
