Apple Launches FastVLM: Ultra-Fast Visual-Linguistic AI Designed for iPhone

Apple introduces FastVLM, its multimodal AI model optimized for iPhone and Mac. Lightweight, fast, and open source, it brings visual-language AI directly to the device.

Read More  
Gemini 2.5 Pro Understands 6 Hours of Video: The New Era of Visual Intelligence

Gemini 2.5 Pro analyzes 6 hours of video, understands events, structure and sequences thanks to a 2 million token window and YouTube parsing. A decisive leap in multimodal AI.

Read More  
GENMO: NVIDIA's New AI Model That Animates the Human Body from Text, Audio, and Video

GENMO is NVIDIA's new AI model capable of estimating and generating human motion from text, audio, video, and keyframes. A generalist framework that revolutionizes the entire field of animation and motion AI.

Read More  
WAN GP AI: The Ultimate Guide to Generate Realistic Video for Free with Skyreels, ControlNet and WAN 2.1 Locally (2025)

WAN GP AI is a free, locally installable AI video suite, including Van 2.1, Skyreels, ControlNet and other models to generate realistic videos with prompts or images, even on mid-range PCs.

Read More  
Google Veo 2 Free on Gemini: Create 16:9 AI Videos Even from Your Smartphone in Just a Few Clicks

Google Veo 2 is now available for free on your smartphone via Gemini. You can generate 16:9 AI videos from text prompts, right from the app, with realistic and stunning results.

Read More  
3DV-TON: The New Standard for Realistic and Consistent AI Try-On Videos

3DV-TON is a new AI model for realistic try-on videos, capable of replacing clothes in videos while maintaining motion, texture and visual fidelity. Ideal for fashion, e-commerce and digital creators.

Read More  
FantasyTalking: The New Frontier of Realistic Talking Portraits

FantasyTalking is an AI model that can realistically animate static faces, synchronizing them with audio to create expressive and coherent talking portraits, perfect for storytelling and interactive applications.

Read More  
EdgeTAM: The New Standard for On-Device AI Tracking, Designed by Facebook Research

EdgeTAM is a lightweight, open-source AI model for real-time tracking, designed for mobile devices. Developed by Facebook Research, it allows you to follow any object in a video directly on-device.

Read More  
SkyReels AI: Create 30-Second Videos or Full-Length Movies with AI

SkyReels AI 2.0 lets you generate 30-second videos or full-length movies from prompts or images. The new system supports storyboards, lip sync, effects, and formats for social or cinema.

Read More  
LemonAI Introduces Slice Live: The Real-Time Audio-Video AI Model That Revolutionizes Digital Interactions

Slice Live by LemonAI ushers in a new era in digital interactions: real-time video calls with lifelike avatars created from a single photo. Discover how it revolutionizes entertainment, education, and the future of virtual communication.

Read More  
MagicLight AI: How to Create Movies, Videos and Cartoons with Artificial Intelligence

MagicLight AI is the AI tool that allows you to create films, shorts and cartoons starting from a simple prompt, revolutionizing the accessibility of animated video production.

Read More  
VIDU Q1 vs KLING 2.0: Which AI Video Generator Is Better (Even Free)?

VIDU Q1 vs KLING 2.0: Two of the best AI video generators out there, also available for free. Find out which one to choose for your creative projects.

Read More