- 3.6 Flash: Our workhorse model that delivers better coding, knowledge work, and multimodal efficiency. According to Artificial Analysis Indexreduces output token usage by 17% compared to Flash 3.5, and in some benchmarks such as DeepSWE, by Data curvewe see up to 65%, all at a lower cost per output token.
- 3.5 Flash version: Our fastest and most cost-effective Class 3.5 model, delivering 350 output tokens per second according to the Artificial Intelligence Index, also significantly outperforming previous generations of Flash-Lite in agentic workflows.
- 3.5 Flash Cyber in CodeMender: Effective cybersecurity applications require careful coordination of the model with the agent infrastructure. We are introducing a combination of a novel, high-performance, specialized cybersecurity-focused model combined with our CodeMender code security agent that delivers competitive performance at the edge.
In addition to today’s releases, Gemini 3.5 Pro is currently in testing with partners and we plan to make it widely available as soon as it is ready. At the same time, our team is already focusing on building novel generation models. We have launched our most ambitious pre-training for Gemini 4 to date and we are thrilled with the progress.
3.6 Flash: More capable and better quality than 3.5 Flash
Gemini 3.6 Flash is based directly on developer and customer feedback on 3.5 Flash. 3.6 Flash not only advances coding and knowledge work, but it does so while significantly improving token performance. For example, in the Artificial Intelligence Index, we see that 3.6 Flash uses 17% fewer output tokens than 3.5 Flash. Implementing multi-step workflows also requires fewer reasoning steps and tool calls.
This increased performance is also combined with a lower price than 3.5 Flash. With $1.50/1M input tokens and $7.50/1M output tokens, Flash 3.6 reduces the overall cost of an agent job, making it more cost-effective to build and run agents.