So no 3.5 Pro then?

Trusted testers are beginning to receive Gemini's desktop Tasks mode, with a Gemini Live entry point and settings for allowed and disallowed apps.

Google’s AI satellite will launch aboard a SpaceX rocket on October 1st.

Would you let Gemini do the talking for you?

The AI avatar can transition between 97 languages.

Gemini 3.8 Live with Live Avatar is now available in Gemini Enterprise, adding a speaking video persona that can keep talking while tools run in the background.

The Google Flow web build now labels a possible Nano Banana update as 2.1, replacing an earlier 2.5 Flash reference. Google has not announced the model.

Gemini 4 Argon begins with trusted cyber defenders. Paid API customers and Google AI Ultra subscribers are first in its planned broader rollout.

Google claims Gemini 4 Argon stacks up against OpenAI and Anthropic’s models.

A Gemini desktop setting suggests work beyond selected folders, internet access, and cross-app actions without approval at every step. It remains in testing.

Google AI Studio Build has an unreleased Ultra mode for advanced skills and tools. Its Argon connection and subscription requirements remain unconfirmed.

Gemini 4 Argon appears to be quietly reaching some users through A/B tests, and the early outputs shared on X look like a clear step up from Gemini 3.8 Flash. Whether it lives up to Google's benchmark claims is another matter.

GemPix 3 Pro has appeared in Google's image citation functionality, hinting at Nano Banana 3 Pro. No public release plans are confirmed.

Personal Intelligence is available in Android Auto starting today.

CarPlay users in the US can now use Gemini to chat with Google Maps.