Google Refreshes Gemini Family: Two New Models Now Available
Google has recently unveiled three innovative artificial intelligence models: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber. This expansion of the Gemini family aims to cater to a broader range of applications, from everyday tasks to specialized cybersecurity challenges. While Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are being rolled out to a wider audience of users and developers, Gemini 3.5 Flash Cyber will be exclusively available to select institutions and trusted partners due to its sensitive capabilities.
Introducing Gemini 3.6 Flash: Enhanced Efficiency and Performance
The standout innovation in this release is the Gemini 3.6 Flash model, engineered for superior performance and cost-effectiveness. This model is specifically designed to excel in tasks such as programming, comprehensive document analysis, intricate chart interpretation, and other processes that demand sophisticated text and image processing capabilities.
Key Advantages of Gemini 3.6 Flash:
- Improved Efficiency: According to tests conducted by Artificial Analysis, Gemini 3.6 Flash consumes 17% fewer output tokens compared to its predecessor, Gemini 3.5 Flash. This translates to more economical operations for AI agents.
- Optimized Reasoning: Google highlights that the model requires fewer reasoning steps and relies less on supplementary tools when executing complex commands, streamlining task completion.
- Superior Performance: It has demonstrated enhanced results in critical areas such as coding, data analysis, and handling computer interfaces. Developers will appreciate its ability to introduce fewer unnecessary code changes and execute multi-stage tasks more efficiently.
Cost Structure for Gemini 3.6 Flash:
The pricing for Gemini 3.6 Flash has also been updated, making it more accessible for large-scale deployments:
- Input Tokens: $1.50 per million tokens
- Output Tokens: $7.50 per million tokens
This revised pricing, combined with the model’s reduced token generation, is expected to significantly lower the operational costs for AI applications. For developers interested in optimizing their AI models, understanding advanced features like Gemini’s memory import can further enhance efficiency.
Gemini 3.5 Flash-Lite: Speed and Versatility for Diverse Tasks
The second model, Gemini 3.5 Flash-Lite, is positioned as a highly efficient and versatile solution, capable of handling a spectrum of tasks from simple to complex. Despite its “Lite” designation, this model is designed with robustness in mind, making it suitable for a wide array of applications.
Applications and Performance:
- Broad Utility: Gemini 3.5 Flash-Lite can be deployed for information retrieval, language translation, document processing, analyzing product databases, and even extracting data from receipts.
- Blazing Speed: Artificial Analysis reports that this model can generate approximately 350 tokens per second, establishing it as the fastest solution within the Gemini 3.5 family.
Flexible Pricing and Customizable Intelligence:
- Input Tokens: $0.30 per million tokens
- Output Tokens: $2.50 per million tokens
A unique feature of Gemini 3.5 Flash-Lite is its customizable “thinking” level. Developers can adjust the model’s processing depth: for simpler tasks, it operates faster and more cost-effectively, while for more demanding processes, it can perform additional reasoning steps to achieve higher accuracy. Intriguingly, Google claims that in certain tests, 3.5 Flash-Lite has even outperformed the larger Gemini 3 Flash model, showcasing its optimized architecture and efficiency. Explore the broader Gemini ecosystem and its applications, including advanced conversational AI, to see how these models integrate.
Gemini 3.5 Flash Cyber: A Specialized Tool for Cybersecurity
The third addition, Gemini 3.5 Flash Cyber, is a highly specialized model dedicated to enhancing cybersecurity measures. It is expertly designed to detect, verify, and remediate vulnerabilities within software code.
CodeMender System Integration:
Gemini 3.5 Flash Cyber operates within the sophisticated CodeMender system, where multiple AI agents can collaboratively analyze the same problem, produce a unified report, and propose effective corrective measures. This multi-agent approach significantly strengthens the defense against cyber threats.
Limited Access for Enhanced Security:
Given its powerful capabilities, which could potentially be misused for both defensive and offensive cyber operations, Gemini 3.5 Flash Cyber will not be publicly released. Instead, its initial deployment will be restricted to government entities and trusted partners participating in a controlled pilot program. This cautious approach ensures the responsible use of such a potent tool.
Availability of the New Gemini Models
Both Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now widely accessible through various platforms:
- Gemini API
- Google AI Studio
- Android Studio
- The official Gemini application
- Enterprise-focused solutions
Additionally, the lighter Gemini 3.5 Flash-Lite model is being progressively integrated into Google Search, extending its utility to everyday search queries.
Frequently Asked Questions (FAQ)
Gemini 3.6 Flash is designed for advanced tasks like programming and complex document analysis, offering higher efficiency with 17% fewer output tokens and streamlined reasoning for complex commands. Gemini 3.5 Flash-Lite, while versatile for a broad range of tasks including information retrieval and translation, focuses on speed, generating approximately 350 tokens per second, making it the fastest in its family, and is more cost-effective for simpler operations.
Gemini 3.5 Flash Cyber is exclusively available to government entities and trusted partners through a limited pilot program. Its availability is restricted due to its advanced capabilities in detecting and remediating software vulnerabilities, which could potentially be exploited for both defensive and offensive cyber operations, necessitating a controlled deployment for responsible use.
Developers can access Gemini 3.6 Flash and Gemini 3.5 Flash-Lite via the Gemini API, Google AI Studio, and Android Studio. These platforms provide the necessary tools and documentation for integrating the new models into various applications and services, enabling a wide range of AI-powered solutions.
For Gemini 3.6 Flash, input tokens cost $1.50 per million, and output tokens cost $7.50 per million. For Gemini 3.5 Flash-Lite, input tokens are $0.30 per million, and output tokens are $2.50 per million. Both models are designed for cost efficiency, with Gemini 3.6 Flash using fewer tokens and Gemini 3.5 Flash-Lite offering very competitive pricing for high-speed processing, aiming to reduce overall operational costs for AI agents.
Source: Google. Opening photo: Google Gemini AI