0
Performance comparison of Gemini 3.5 Flash and the new Gemini 3.6 Flash / Courtesy of Google

Google has introduced its next-generation AI models focused on enhancing operational efficiency and lowering token usage for complex agentic workflows: Gemini 3.6 Flash and Gemini 3.5 Flash-Lite.

According to Google on July 21 (local time), the newly announced Flash family is designed to optimize AI agent development by delivering both stronger capabilities and superior cost-effectiveness. The trio includes Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and the security-specialized Gemini 3.5 Flash Cyber.

Gemini 3.6 Flash improves coding, knowledge work, and multimodal reasoning capabilities over Gemini 3.5 Flash while reducing average output token consumption by 17 percent. On specific benchmarks such as Datacurve's DeepSWE, token reductions reach up to 65 percent. By minimizing intermediate reasoning steps and tool calls, the model significantly trims overall operational expenses for multi-step AI workflows.

Pricing has been updated to $1.50 per 1 million input tokens and $7.50 per 1 million output tokens, giving it a sharper competitive edge.

In terms of benchmark performance, Gemini 3.6 Flash recorded notable improvements across code modification accuracy, machine learning research, computer use, document parsing, and report generation. Enterprise clients such as Heavia and Harvey reported boosted performance in multimodal tasks involving visual processing, chart analysis, and data extraction.

Alongside 3.6 Flash, Google released Gemini 3.5 Flash-Lite, targeted at high-throughput agentic search and large-scale document processing tasks. Generating up to 350 output tokens per second, it stands as the fastest model in the 3.5 series. At $0.30 per 1 million input tokens and $2.50 per 1 million output tokens, it provides a cost-efficient solution for volume-heavy enterprise operations.

Meanwhile, Gemini 3.5 Flash Cyber is tailored for detecting and remediating software vulnerabilities in integration with CodeMender, Google's automated code repair agent. This specialized model will be made available through a restricted preview to government agencies and trusted security partners.

Google is rolling out Gemini 3.6 Flash and Gemini 3.5 Flash-Lite sequentially starting today across Google AI Studio, Gemini API, Android Studio, Gemini Enterprise, and the consumer Gemini app.

                                                                                                           Jeong A-reum