Google just dropped three things at once. A new voice model. A global expansion of Search Live. And their cheapest model ever. All in one day. Gemini 3.1 Flash Live is here and it changes how you talk to Google.

Gemini 3.1 Flash Live is Google's new audio and voice model. They're calling it their highest-quality voice model ever. And based on what they shipped today, that's not marketing fluff.

The model powers Gemini Live and Search Live. Responses are faster with fewer awkward pauses. It can follow your conversation for twice as long as before, meaning longer brainstorms without losing context. It understands pitch, tone, and emphasis better than any previous version. And it filters out background noise so it actually works in the real world, not just in a quiet demo room.

It supports over 90 languages for real-time multimodal conversations. No switching settings. Just talk in your language and it works.

For developers, 3.1 Flash Live is available today via the Gemini Live API in Google AI Studio. You can build voice-first agents that complete complex tasks in real time. Home Depot and Verizon are already using it in their customer experience systems.

All audio generated by the model is watermarked with SynthID so you can detect when audio is AI-generated. That matters as voice AI gets harder to distinguish from real humans.

This is the part that affects everyone. Search Live is now available in over 200 countries and territories. Anywhere AI Mode is active, you can now have a real-time voice and camera conversation with Google Search.

That means you can talk to Google instead of typing. Point your camera at something and ask what it is. Have a back-and-forth conversation about what you're looking at. Get web links alongside spoken responses. Ask follow-up questions naturally without starting over.

This is Google's biggest bet on the future of search. Typing keywords is starting to feel outdated when you can just talk or show your phone something and get instant answers. If this takes off at scale, it changes how billions of people find information online.

Google also dropped Gemini 3.1 Flash-Lite recently. Their most cost-efficient model ever at $0.25 per million input tokens. That's 8x cheaper than Gemini 3.1 Pro. It's 2.5x faster to first token than 2.5 Flash with 45% faster output speed. On benchmarks it beats GPT-5 mini and Claude 4.5 Haiku across multiple tests.

But the real story isn't any single model. It's the pace. In March alone Google shipped Gemini 3.1 Pro, Gemini 3.1 Flash-Lite, Gemini 3.1 Flash Live, Search Live in 200+ countries, Stitch 2.0 with voice design, and Gemini inside Google Maps. That's six major drops in one month.

Google isn't just keeping up anymore. They're setting the pace. Faster models. Cheaper models. Voice-first experiences. Global rollouts. While everyone else is focused on agents and coding tools, Google is quietly rebuilding every product billions of people already use. Search. Maps. Gmail. Docs. All of it is becoming Gemini-native.

That's the advantage nobody talks about. OpenAI has the brand. Anthropic has the developers. But Google has the distribution. And they're using every inch of it.

OpenAI quietly shelved its planned erotic chatbot. The company had been working on a consumer-facing AI product for romantic and sexual conversations. But employees and investors pushed back hard over concerns about the broader social effects of sexualized AI content. OpenAI pulled the plug. In a market where every company is racing to ship faster, OpenAI chose to slow down on this one. Sometimes knowing what not to build matters more than what you build.

2x longer conversations. Fewer awkward pauses. Search Live is now in 200+ countries so you can talk to Google instead of typing. OpenAI killed their erotic chatbot before it shipped. And Google isn't slowing down. Six major AI drops in March alone. They're not chasing the competition anymore. They are the competition.

Stay building. 🤖