Google DeepMind has released Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber, expanding its selection of lower-cost models for coding, AI agents and cybersecurity work. The company focused the release on speed, token efficiency and reliability rather than introducing a new Gemini Pro model.
Gemini 3.6 Flash is now priced at $1.50 per million input tokens and $7.50 per million output tokens. Google describes it as a general-purpose model for agentic workflows, multimodal tasks and software development.
Gemini 3.6 Flash Uses Fewer Tokens
According to Google’s official model announcement, Gemini 3.6 Flash improves coding, knowledge work and multimodal performance compared with Gemini 3.5 Flash. It used 17% fewer output tokens during testing on the Artificial Analysis Index and required fewer reasoning steps and tool calls for multi-stage tasks.
Google said the efficiency improvements could lower the cost of running AI agents that repeatedly call tools or generate long responses. The model is available through the Gemini API and Google’s enterprise AI services.
Gemini 3.5 Flash-Lite is designed for applications that process large volumes of requests at lower cost. Google charges $0.30 per million input tokens and $2.50 per million output tokens, making it the least expensive model in the Gemini 3.5 family.
Cyber Model Targets Software Vulnerabilities
Gemini 3.5 Flash Cyber is a specialised version trained to find, confirm and repair security vulnerabilities. Google said the model can be used for frequent code scans, software-release checks and automated vulnerability research.
The model will initially be offered through a limited-access pilot for governments and selected partners. Google said it is already being used internally across products including Chrome, Android, Cloud, Ads and YouTube.
During one evaluation, Gemini 3.5 Flash Cyber found 55 confirmed issues in the V8 JavaScript engine. Google reported that Gemini 3.5 Flash found 47, while Claude Opus 4.6 found 36 under the same testing conditions.
Gemini Pro Update Remains Pending
The release did not include the expected successor to Google’s Gemini Pro model, which is intended for more demanding reasoning and coding tasks. Google previously said Gemini 3.5 Pro was being used internally and was expected to roll out after the May release of Gemini 3.5 Flash.
Google DeepMind product lead Logan Kilpatrick said the company is testing the Pro model with partners. He also said work has begun on the company’s largest pre-training run for Gemini 4.
Featured image credits: Carlos Luna via Flickr
For more stories like it, click the +Follow button at the top of this page to follow us.