Droplet3D Commonsense Priors from Videos Facilitate 3D Generation
MULTIMODAL AI
-
AI Anime Created With Nano Banana and Seedance 1.0
By
–
Impressionnant ! Un anime IA crée avec Nano Banana et Seedance 1.0. C’est simple, vraiment pas cher !
— VISION IA (@vision_ia) 1 septembre 2025
Ce n'est que la version la plus mauvaise de l'IA que nous n’aurons jamais. On ce rapproche en plus de l'épisode en 1 prompt ! https://t.co/1l2gSf0kzx pic.twitter.com/fXNKPMCUPxImpressionnant ! Un anime IA crée avec Nano Banana et Seedance 1.0. C’est simple, vraiment pas cher ! Ce n'est que la version la plus mauvaise de l'IA que nous n’aurons jamais. On ce rapproche en plus de l'épisode en 1 prompt !
-
xAI Releases All Major Grok Features Free This Month
By
–
Toutes les sorties de xAI ce mois-ci étaient entièrement GRATUITES :
Grok Imagine… génération d’images et de vidéos
Grok Code Fast-1
Grok 2.5 en open source
Améliorations du traitement PDF de Grok -
ByteDance Releases USO Open Style-Subject Optimized Model
By
–
ByteDance released USO, an open ‘style-subject optimized customization model’
— The Rundown AI (@TheRundownAI) 1 septembre 2025
It can preserve subjects and apply new artistic styles to create customized images
Available on Hugging Facepic.twitter.com/KNT359g98xByteDance released USO, an open ‘style-subject optimized customization model’ It can preserve subjects and apply new artistic styles to create customized images Available on Hugging Face
-
Higgsfield Speak 2.0 Launches Realistic AI Avatar Video Tool
By
–
Higgsfield AI launched Higgsfield Speak 2.0, a new upgrade to its custom avatar tool It brings more realistic motion, advanced lip-sync, and enhanced video control Also supports 70+ languages https://
x.com/higgsfield_ai/
status/1961525851352432829/video/1
… -
R-4B: Auto-Thinking MLLMs via Bi-Mode Annealing and RL
By
–
R-4B Incentivizing General-Purpose Auto-Thinking Capability in MLLMs via Bi-Mode Annealing and Reinforce Learning
-
EmbodiedOneVision: Vision-Text-Action Pretraining for Robot Control
By
–
EmbodiedOneVision Interleaved Vision-Text-Action Pretraining for General Robot Control
-
Image Generation Flaws Need Comprehensive Benchmark Testing
By
–
Yes, that makes sense. Image generation is deeply flawed today. Would love to see a range of these questions used across a variety of imagegen and done over time. Would be an interesting benchmark.
-
Veo 3 Google AI Model Generates Optimal Results with Simple Prompts
By
–
Finalmente el ganador para mi es Veo 3, que también tenemos disponible en Freepik y con el que he conseguido el resultado óptimo!
— Carlos Santana (@DotCSV) 31 août 2025
Un simple prompt donde explico que tiene que aparecer un pangolín a mi lado y el modelo de Google hace su magia ✨ pic.twitter.com/mIuXFKOcbFFinalmente el ganador para mi es Veo 3, que también tenemos disponible en Freepik y con el que he conseguido el resultado óptimo! Un simple prompt donde explico que tiene que aparecer un pangolín a mi lado y el modelo de Google hace su magia