Google said the model completes multi-step workflows using fewer reasoning steps and tool calls than its predecessor. Across several benchmarks, the company reported improvements in coding, computer use and knowledge work, including higher scores on DeepSWE, MLE Bench, OSWorld-Verified and GDPval-AA v2. Google also highlighted customer use cases involving document parsing, chart analysis, report drafting and financial data analysis.
The company said Gemini 3.6 Flash includes enhanced Frontier Safety protections for chemical, biological, radiological, nuclear and cyber misuse scenarios. According to Google, the safeguards make the model more resistant to jailbreak attempts while reducing unnecessary refusals for legitimate use cases.
Google also released Gemini 3.5 Flash-Lite, describing it as the fastest and most cost-effective model in the Gemini 3.5 family. The company said Artificial Analysis measured the model at 350 output tokens per second. Priced at $0.30 per million input tokens and $2.50 per million output tokens, Google said Flash-Lite is designed for high-throughput workloads such as agentic search and document processing.
According to Google, Gemini 3.5 Flash-Lite delivers significant gains over Gemini 3.1 Flash-Lite in coding, long-context understanding and real-world task execution. The company also said the model outperformed Gemini 3 Flash on several agentic and coding evaluations, including SWE-Bench Pro and OSWorld-Verified.
The third release, Gemini 3.5 Flash Cyber, is a specialized version of Gemini 3.5 Flash that has been fine-tuned to identify and remediate cybersecurity vulnerabilities. Google said the model operates within its CodeMender security agent, where multiple Flash Cyber agents work together to produce a combined report. According to the company, the system achieves competitive performance on the CyberGym benchmark while offering a lower price per token than larger models.
Because of the model's cybersecurity capabilities, Google said Gemini 3.5 Flash Cyber will initially be available only to governments and trusted partners through a limited-access pilot program.
Alongside the new model launches, Google provided an update on its flagship Gemini Pro family. The company said Gemini 3.5 Pro is currently being tested with partners and will become broadly available when it is ready. Google also disclosed that it has begun what it described as its most ambitious pre-training run yet for Gemini 4.
Google said Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are available starting today through the Gemini API in Google AI Studio and Android Studio, with Gemini 3.6 Flash also available in Google Antigravity. Enterprise customers can access the models through the Gemini Enterprise Agent Platform, while Gemini 3.6 Flash is also available in the Gemini Enterprise app. Google said Gemini 3.5 Flash-Lite is rolling out in the Gemini app and Google Search.
This analysis is based on reporting from Google.
Image courtesy of Google.
This article was generated with AI assistance and reviewed for accuracy and quality.
