Google Unveils Gemini 3.1 Flash Live

Welcome to another edition of AI PlanetX.

Google unveils Gemini 3.1 Flash Live; Anthropic reveals its most powerful model; Meta’s TRIBE v2 beats real brain data.

• Hottest AI News

• Top AI & SaaS Tools

• AI Tutorial: Create Explorable 3D World

• Top AI & Tech News

• AI Art Spotlight

• Prompt of the Day: OpenAI Prompt Guide for GPT-5.4

• Featured AI Video

Hottest AI News

DeepMind

Google Unveils Gemini 3.1 Flash Live as Its Most Natural Voice AI Yet

Google has launched Gemini 3.1 Flash Live, its most natural voice AI yet. It is rolling out in 200+ countries for Gemini Live and Search Live, and is also in developer preview via the Gemini Live API in Google AI Studio.

Details:

• It supports real-time, multimodal conversations in 90+ languages. The model can detect pitch, pace, and emotion while filtering noisy backgrounds like traffic or loud TVs

• It reduces awkward pauses with ultra-low latency, reaching 0.96-second response times on its Minimal setting. It also doubles memory, adjusts tone dynamically, and scores 95.9% on Big Bench Audio

• Developers get better guardrail adherence and more reliable external tool use. Pricing stays the same as before at $0.35 per hour for audio input and $1.40 per hour for output

Gemini 3.1 Flash Live blends natural conversation, strong reasoning, and developer-friendly tools to set a higher standard for real-time voice agents.

Become the go-to AI expert in 30 days

AI keeps coming up at work, but you still don't get it?

That's exactly why 1M+ professionals working at Google, Meta, and OpenAI read Superhuman AI daily.

Here's what you get:

• Daily AI news that matters for your career - Filtered from 1000s of sources so you know what affects your industry.

• Step-by-step tutorials you can use immediately - Real prompts and workflows that solve actual business problems.

• New AI tools tested and reviewed - We try everything to deliver tools that drive real results.

• All in just 3 minutes a day

Join 1M+ pros

Anthropic

Anthropic Mistakenly Reveals Its Most Powerful AI Model

A CMS error at Anthropic made nearly 3,000 unpublished files public, revealing an unreleased AI model, serious security warnings, and private company plans before the database was secured.

Details:

• The published drafts mention a model called Claude Mythos. It includes a top tier named Capybara, described as larger, smarter, and costlier than Opus 4.6, with major gains in coding, reasoning, and cybersecurity

• Anthropic sees the system as far ahead in cyber ability and is worried about hacking risks. Early use is being limited to cyber defenders, similar to the recent high-risk treatment of GPT-5.3-Codex

• The data trove was found by Roy Paz and Alexandre Pauwels. It also included parental leave details and plans for a private English manor retreat with CEO Dario Amodei and European leaders

Anthropic locked down the search and blamed an external tool, but the incident still reveals stark dual‑use cyber risks and aggressive expansion in frontier AI.

Top AI & SaaS Tools

• Robomotion (Life-time Deal): RPA platform that uses AI drag-and-drop bots to automate repetitive web and desktop tasks across Mac, Windows

• Voxtral TTS: Mistral AI's TTS delivers low-latency, emotionally expressive, multilingual speech with zero-shot voice cloning and easy integration into STT+LLM workflows

• Codex Plugins: OpenAI announced that Codex currently supports plugins, integrating with popular tools (Slack, Figma, Notion, Gmail, Google Drive, Docs, Sheets, Slides, etc.)

• Wispr Flow for Android: Turns messy speech into clear, well‑punctuated text in any app by removing filler words and self‑corrections [F-R-E-E & Unlimited]

• Cohere Transcribe: Open-source speech-to-text model that achieved top English accuracy on HF’s Open ASR leaderboard with a 5.42% word error rate [F-R-E-E]

Stop Drowning In AI Information Overload

Your inbox is flooded with newsletters. Your feed is chaos. Somewhere in that noise are the insights that could transform your work—but who has time to find them?

The Deep View solves this. We read everything, analyze what matters, and deliver only the intelligence you need. No duplicate stories, no filler content, no wasted time. Just the essential AI developments that impact your industry, explained clearly and concisely.

Replace hours of scattered reading with five focused minutes. While others scramble to keep up, you'll stay ahead of developments that matter. 600,000+ professionals at top companies have already made this switch.

Join them today, for free.

AI Tutorial

Create Fully Explorable 3D World from Prompt

Creating a 3D world no longer takes hours of setup and design work. With OpenArt Worlds, you can turn an image or prompt into a full 3D environment, then explore it, customize it, and capture polished visuals.

1. Launch OpenArt Worlds

   Go to OpenArt Worlds and sign in to your account.

2. Start Fresh 3D World

   Tap on "Create 3D World" to begin the setup.

3. Choose Input Method

   You can start in one of two ways:

   • Upload an image

   • Enter a text prompt

4. Write Strong Scene Prompt

   If you choose the prompt option, be very specific about depth, perspective, and structure.

   Example prompt:

   A detailed cyberpunk skyline with layered buildings, bright neon signs, deep street perspective, atmospheric fog, dramatic lighting, realistic shadows, cinematic composition, wide camera view, high-detail 3D environment

5. Generate Environment

   Once your image or prompt is processed, the tool creates a full 3D world that you can move through and explore.

6. Add Elements to Scene

   After the world is generated, you can place extra elements where needed, such as:

   • Characters

   • Objects

   • Scene details

7. Refine Camera and Capture Visuals

   Adjust the camera angles, frame the exact shots you want, and capture production-ready visuals directly inside the environment.

Note: Focus on the overall structure and depth first, then refine the smaller scene details once the base environment looks strong.

Top AI & Tech News

• Meta open-sourced TRIBE v2, an AI trained on brain scans from 700+ people that simulates vision, hearing, and language—and its synthetic predictions beat real fMRI data

• Apple plans to expand Siri to other AI models starting with iOS 27, ending ChatGPT’s sole role and letting users choose which AI handles their requests

• OpenAI has indefinitely paused development of an “adult mode” (a mature-content feature) for ChatGPT after internal and external criticism and repeated delays

• Wikipedia’s volunteer editors banned AI-written articles on the English site, calling it a pushback against “enshittification” and aggressive AI promotion by companies

AI Art Spotlight

Create Your Own Photo Like This Using the Prompt Below:

Cinematic studio portrait of a young man (use the uploaded image as a reference for the face) with messy, textured hair, wearing a black turtleneck. Centered composition with a dark, minimal background. Dramatic low-key lighting, a horizontal red neon light strip cutting across his eyes, and strong contrast shadows covering the lower part of the face. Intense gaze, moody atmosphere, cyberpunk aesthetic. Ultra-realistic, sharp focus, 85mm lens, shallow depth of field, high detail, editorial photography style.

Model: Nano Banana 2

Prompt of the Day

Prompt Guidance for GPT-5.4 by OpenAI

This guide shows how to write better prompts for GPT‑5.4, helping developers structure clear instructions for more accurate and consistent results. It is part of the OpenAI API documentation and focuses on recommended prompting patterns and common mistakes.

Featured AI Video

How to Run Open-Source AI Models