Google has introduced two new additions to its Gemini family of artificial intelligence models, Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, as the company looks to strengthen its offerings for developers building AI agents and enterprise applications. The new models focus on improving speed, efficiency and cost effectiveness while supporting increasingly complex AI workflows.
The launch comes as technology companies race to optimise AI models for production environments, where businesses are seeking lower operating costs, faster response times and improved reliability. Google said the latest Flash models have been designed to address these requirements while maintaining strong performance across coding, reasoning and multimodal tasks.
Gemini 3.6 Flash succeeds Gemini 3.5 Flash as Google's flagship efficiency-focused model. According to the company, it delivers stronger performance in coding, knowledge work and multimodal applications while consuming significantly fewer output tokens. Google said the model requires fewer reasoning steps and tool calls to complete complex workflows, making AI agents faster and less expensive to operate. Pricing has also been reduced compared with Gemini 3.5 Flash, lowering the cost of deploying agentic AI applications at scale.
Alongside it, Google launched Gemini 3.5 Flash-Lite, a lightweight model aimed at high-volume AI workloads. The model has been designed for applications where low latency and affordability are priorities, including document parsing, data extraction and sub-agent tasks that require rapid responses. It supports text, images, video, audio and PDF inputs, allowing developers to deploy multimodal AI experiences without significantly increasing infrastructure costs.
Both models are intended to support the growing adoption of AI agents, software systems capable of executing multi-step tasks with limited human intervention. Google said developers building production-grade AI applications increasingly require models that balance intelligence with operational efficiency, particularly as enterprise deployments scale.
The announcement also included an update on Google's broader AI roadmap. The company confirmed that training has begun for Gemini 4, its next-generation frontier model, while Gemini 3.5 Pro remains in testing with selected partners. Google said the model will be made broadly available after completing additional validation and performance testing.
The latest releases arrive at a time of heightened competition in the generative AI market, where companies including OpenAI, Anthropic and Meta continue to introduce increasingly capable foundation models. Rather than focusing solely on raw performance, technology providers are also competing on efficiency, inference costs and the ability to power enterprise-grade AI agents across industries.
Google said Gemini 3.6 Flash is particularly suited for coding, agentic execution and spatial reasoning, while Gemini 3.5 Flash-Lite is designed for large-scale production environments where throughput and response speed are critical. Both models are now available through Google AI Studio and the Gemini API, enabling developers and enterprises to begin integrating them into applications.
The launch reinforces Google's strategy of expanding the Gemini portfolio with specialised models tailored for different workloads instead of relying on a single general-purpose AI system. As enterprise adoption of AI agents accelerates, efficiency-focused models are expected to play a larger role in reducing deployment costs while supporting more complex business processes.
Related Articles
