Google has launched the Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber models, focusing on efficiency and cybersecurity enhancements.
Google has launched three new models in its Gemini Flash series: the Gemini 3.6 Flash, 3.5 Flash-Lite, and the specialized 3.5 Flash Cyber. The 3.6 Flash model will be sold at a lower price than the previous version, while access to the Cyber variant is initially restricted to a pilot program for select government and trusted partners.
Enhanced Performance and Efficiency
The Gemini 3.6 Flash is positioned as Google's main model for coding, data processing, and multi-faceted tasks. The company claims this version is 17% more efficient in token use compared to the 3.5 Flash, with certain applications seeing reductions in output tokens by as much as 65% during specific tests such as DeepSWE. Pricing is set at $1.50 per million input tokens and $7.50 for output tokens.
Benchmark tests revealed significant performance enhancements:
- DeepSWE scores improved to 49%, up from 37% for the previous model;
- MLE Bench results reached 63.9%, compared to 49.7%;
- OSWorld-Verified tests showed an increase to 83.0% from 78.4%.
These advancements have led to greater usability, confirmed by the OSWorld benchmarks. The model is accessible through an API and the Gemini Enterprise application.
Fast and Cost-Effective Options
The Gemini 3.5 Flash-Lite is touted as the quickest and most economical option in the lineup, achieving a speed of 350 output tokens per second, with pricing at $0.30 per million input tokens and $2.50 for output tokens. It reportedly outperforms its predecessor, the 3.1 Flash-Lite, in agent scenarios and shows better results in tests including SWE-Bench Pro and OSWorld-Verified.
Focus on Cybersecurity
The Gemini 3.5 Flash Cyber is specifically designed for detecting and addressing cybersecurity vulnerabilities. It employs a team of agents to generate comprehensive reports as part of the CodeMender initiative. Its performance on the CyberGym benchmark is competitive, which is why its access is currently limited to governmental and trusted entities during the pilot phase.
Developers can currently access the 3.6 Flash and 3.5 Flash-Lite models through the Gemini API in Google AI Studio and Android Studio. Businesses can utilize these models through the Gemini Enterprise Agent Platform, with the 3.6 Flash also integrated into Gemini Enterprise. Regular users are able to access them via the Gemini app, with the rollout of the 3.5 Flash-Lite in Google Search already underway. Meanwhile, partners are testing the flagship Gemini 3.5 Pro, and Gemini 4 is in the early stages of training.
Additionally, Google announced the introduction of the new Frozen v2 chip, which includes integrated Gemini architecture.
