A Complete Guide to Google's Gemini 3.6 Flash, 3.5 Flash-Lite, and Cyber
Discover Google's latest AI models: Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. Learn about their pricing, token efficiency, and agentic workflows.
Author
Shalimar Mehra
Gemini 3.6 Flash: The Efficient Workhorse
Gemini 3.6 Flash builds on the foundation of 3.5 Flash, bringing substantial improvements in coding, knowledge work, and multimodal capabilities.
Superior Token Efficiency
One of the most notable upgrades is its token efficiency. According to the Artificial Analysis Index, 3.6 Flash consumes 17% fewer output tokens than its predecessor. In specific coding benchmarks like Datacurve's DeepSWE, token reduction reaches up to 65%. Because it requires fewer reasoning steps and tool calls to complete multi-step tasks, it actively drives down the overall cost of agentic workflows.
Benchmark Performance
Despite its efficiency, 3.6 Flash achieves significant performance gains:
Coding: Achieves 49% (up from 37%) on DeepSWE, exhibiting higher precision and fewer unwanted code edits.
Machine Learning Research: Scores 63.9% (up from 49.7%) on MLE Bench.
Computer Use: Improves to 83.0% (from 78.4%) on OSWorld-Verified.
Knowledge Work: Reaches a score of 1421 (up from 1349) on GDPval-AA v2, proving highly capable of parsing documents and analyzing charts.
Enhanced Safety Safeguards
3.6 Flash includes robust Frontier Safety safeguards specifically guarding against Chemical, Biological, Radiological, and Nuclear (CBRN) threats, as well as cyber offense misuse. These protocols make the model highly resistant to jailbreaking while ensuring it minimizes refusals for legitimate, beneficial requests.
Gemini 3.5 Flash-Lite: Built for Speed and Scale
For tasks where raw throughput and ultra-low latency are paramount, Google introduced Gemini 3.5 Flash-Lite.
Speed and Cost
Clocking in at 350 output tokens per second, 3.5 Flash-Lite is the fastest model in the Gemini 3.5 family. It offers an exceptional price-to-performance ratio for businesses managing high-volume production traffic.
Outperforming Heavier Models
Remarkably, 3.5 Flash-Lite frequently outperforms even the Gemini 3 Flash model in agentic and coding evaluations, including SWE-Bench Pro (54.2% vs. 49.6%) and OSWorld-Verified (74.0% vs. 65.1%). It is highly configurable; developers can optimize it for low-latency basic execution or engage higher "thinking levels" for complex subagent workloads.
Gemini 3.5 Flash Cyber: Frontier Security
AI models are identifying security vulnerabilities faster than human teams can patch them. Gemini 3.5 Flash Cyber addresses this by integrating with Google's CodeMender agent infrastructure.
Performance: Multiple 3.5 Flash Cyber agents work together to generate combined vulnerability reports, achieving competitive frontier performance on the CyberGym benchmark.
Availability: Due to the dual-use nature of cyber-offensive AI capabilities, this model is currently restricted to governments and trusted partners in a limited-access pilot program.
Model Comparison Table
Feature | Gemini 3.6 Flash | Gemini 3.5 Flash-Lite |
|---|---|---|
Input Pricing (per 1M tokens) | $1.50 | $0.30 |
Output Pricing (per 1M tokens) | $7.50 | $2.50 |
Key Strength | Multi-step workflows, complex coding, multimodal parsing | Ultra-low latency, high-volume throughput |
Speed (Tokens/sec) | Highly efficient (up to 65% fewer tokens in some tasks) | 350 output tokens/sec |
Target Use Case | Deep reasoning, master agent orchestration | Agentic search, document processing, subagent tasks |
Data sourced directly from Google's product announcement.
Step-by-Step Guide: How to Access the New Models
Developers and enterprises can start using Gemini 3.6 Flash and 3.5 Flash-Lite immediately:
For Developers: Access the models via the Gemini API in Google AI Studio, Android Studio, and Google Antigravity.
For Enterprises: Log into the Gemini Enterprise Agent Platform or the Gemini Enterprise app.
For General Users: Interact with the models via the consumer Gemini app. Gemini 3.5 Flash-Lite is also being rolled out in Google Search
Frequently Asked Questions
How much does Gemini 3.6 Flash cost? Gemini 3.6 Flash is priced at $1.50 per 1 million input tokens and $7.50 per 1 million output tokens, which is lower than the pricing for 3.5 Flash.
What makes Gemini 3.6 Flash different from 3.5 Flash? Gemini 3.6 Flash offers up to 17% lower output token usage (and up to 65% on specific coding benchmarks), making it cheaper and less verbose while improving performance in coding, math, and knowledge work.
Is Gemini 3.5 Flash Cyber available to the public? No. Because of the security implications of AI vulnerability detection, Gemini 3.5 Flash Cyber (via CodeMender) is exclusively available to governments and trusted partners in a limited pilot program.
What is the next Gemini model? Alongside the release of 3.6 Flash and 3.5 Flash-Lite, Google announced they are currently testing Gemini 3.5 Pro with partners, and have already begun the ambitious pre-training run for Gemini 4.
All the Images i used in this Article is from Google Official blog - Google Blog
Google Changes the Name - Gemini Notebook (Formerly NotebookLM): The Ultimate Guide to Pricing, Features, & Privacy
Discussion
Loading comments...
Join the discussion
Log in to share your thoughts and interact with other developers and readers
Log in to Comment