How to Install / Upgrade: Google announces Gemini 3.6 Flash and cybersecurity AI, teases 3.5 Pro and Gemini 4

Google announced Gemini 3.6 Flash and a cybersecurity AI, along with new 3.6 and 3.5 models. The company also teased 3.5 Pro and Gemini 4, and is already training Gemini 4. That is the full scope of this release as reported by Ars Technica.

To install or upgrade, open the Google product or API surface where you already select Gemini models and look for the new 3.6 and 3.5 model names, including Gemini 3.6 Flash. Point your existing apps, agents, or chat sessions at those models in place of the ones you use now. Treat 3.5 Pro as announced only when it actually appears in that same model list, and do not try to select Gemini 4 yet because it is still being trained. If you use a cybersecurity workflow that previously relied on other Gemini models, switch it to the new cybersecurity AI option once it shows up in the same product.

Gotchas: Gemini 4 is not a shipping model in this announcement, only a model Google is training, so any install path that expects Gemini 4 will fail or be unavailable. 3.5 Pro is teased rather than confirmed as fully available, so do not hard-code it until you can select it. Verify the upgrade by confirming the selected model name is Gemini 3.6 Flash or one of the new 3.5 models, and that any cybersecurity path is using the new cybersecurity AI rather than an older model.