Google just launched Gemini 3.1 Flash Live, a new realtime model built for voice and vision agents. After a year of work on model quality, infrastructure, and UX, they're calling it a step-function improvement in quality, reliability, and latency. The race to build the best
MULTIMODAL AI
-
Gemini 3.1 Flash Live launches for real-time voice vision agents
By
–
Listen up 🔊Gemini 3.1 Flash Live is launching today, making a big difference for developers who are building real-time voice and vision agents. How, you ask? Well, this model delivers: — Responses that feel as fast as natural dialogue — Better task completion in noisy environments — Improvements in complex-instruction following
-
Gemini 3.1 Flash Live: High-Quality Audio Voice Model Launched
By
–
Gemini 3.1 Flash Live, our highest-quality audio and voice model, is launching today! This is how it advances our real-time dialogue capabilities:
— Google AI (@GoogleAI) 26 mars 2026
— Faster: 3.1 Flash Live powers faster responses than the previous model, which is great for when you need a timely answer (Ex: “Can… pic.twitter.com/adaT3xB2LXGemini 3.1 Flash Live, our highest-quality audio and voice model, is launching today! This is how it advances our real-time dialogue capabilities: — Faster: 3.1 Flash Live powers faster responses than the previous model, which is great for when you need a timely answer (Ex: "Can you help me change this tire in under 5 minutes?!?") — Longer: In Gemini Live, the model's context window is now twice as long as before, so it can keep up with all of the details shared in your conversations (Ex: "I'm back to writing my future bestselling crime novel. Remind me, who is the secret double agent?") — Global: 200+ more regions will be able to have real-time, multimodal conversations in their preferred language [Translated from EN to English]
-
Google Launches Gemini 3.1 Flash Live Realtime Model
By
–
Introducing Gemini 3.1 Flash Live, our new realtime model to build voice and vision agents!! We have spent more than a year improving the model + infra + experience, the results? A step function improvement in quality, reliability, and latency.
→ View original post on X — @officiallogank, 2026-03-26 15:19 UTC
-
Voxtral TTS: Multilingual Voice Synthesis for Professional Applications
By
–
Voxtral TTS is built for global applications supporting 9 languages and powering voice workflows.
— Mistral AI (@MistralAI) 26 mars 2026
✅ Full audio intelligence: Works with Voxtral Transcribe for end-to-end speech-to-speech, or plugs into any STT + LLM stack.
✅ Built for business: From customer support to… pic.twitter.com/NXTB9KrtrkVoxtral TTS is built for global applications supporting 9 languages and powering voice workflows. ✅ Full audio intelligence: Works with Voxtral Transcribe for end-to-end speech-to-speech, or plugs into any STT + LLM stack. ✅ Built for business: From customer support to real-time translation, it's the output layer that passes the human test. 🎥 See it in action: [Translated from EN to English]
→ View original post on X — @mistralai, 2026-03-26 15:00 UTC
-
Mistral Introduces Voxtral TTS, Its Revolutionary Speech Synthesis Model
By
–
🔊Introducing Voxtral TTS: our new frontier open-weight model for natural, expressive, and ultra-fast text-to-speech
— Mistral AI (@MistralAI) 26 mars 2026
🎭Realistic, emotionally expressive speech.
🌍Supports 9 languages and accurately captures diverse dialects.
⚡Very low latency for time-to-first-audio.
🔄Easily… pic.twitter.com/Q2mdo8UBVo🔊Introducing Voxtral TTS: our new frontier open-weight model for natural, expressive, and ultra-fast text-to-speech 🎭Realistic, emotionally expressive speech.
🌍Supports 9 languages and accurately captures diverse dialects.
⚡Very low latency for time-to-first-audio.
🔄Easily adaptable to new voices [Translated from EN to English]→ View original post on X — @mistralai, 2026-03-26 15:00 UTC
-
New Video Content from Mistral AI
By
–
— Mistral AI for Developers (@MistralDevs) 26 mars 2026
→ View original post on X — @mistralai, 2026-03-26 14:54 UTC
-
Meta Launches TRIBE v2 Foundation Model for Brain Activity Prediction
By
–
Meta is back: Meta just dropped TRIBE v2, a foundation model that predicts how your brain responds to sight, sound, and language.
— Chubby♨️ (@kimmonismus) 26 mars 2026
Trained on 500+ hours of fMRI data from 700+ people, it can predict a new person's brain activity without any retraining, and its predictions are… https://t.co/kjdqTadDiC pic.twitter.com/jmvN4XTYiYMeta is back: Meta just dropped TRIBE v2, a foundation model that predicts how your brain responds to sight, sound, and language. Trained on 500+ hours of fMRI data from 700+ people, it can predict a new person's brain activity without any retraining, and its predictions are
-
AI Generates Cinematic Fantasy Film Scene with Creative Prompting
By
–
Prompt: A cinematic feature film about humanoid-toad wearing a wide brimmed hat and a long cloak visits an old hag to get medicine from her potion shop in a foggy marsh in the swamp. pic.twitter.com/7lOw12gwNj
— Runway (@runwayml) 26 mars 2026Prompt: A cinematic feature film about humanoid-toad wearing a wide brimmed hat and a long cloak visits an old hag to get medicine from her potion shop in a foggy marsh in the swamp.