List of AI News about world model
| Time | Details |
|---|---|
|
2026-10-19 17:36 UTC |
Viggle releases JST-2, a 3D physics-native world model
According to ViggleAI via X, the company released JST-2, a full-stack physics-native world foundation model that builds in 3D rather than pixels and generates expandable 3D scenes and motion-ready characters in seconds. |
|
2026-09-29 11:12 UTC |
DexTacWAM Sets 70.6 Benchmark in Tactile Robotics
According to @berkeley_ai, DexTacWAM tops six dexterous tasks with 70.6 average, outperforming 38.0 baseline via visuo tactile world modeling. |
|
2026-09-15 11:07 UTC |
Odyssey3 Unifies control across robots and cars
According to The Rundown AI, Odyssey-3 controls arms, humanoids, cars, drones, and games, learning new machines in hours from real-world video. |
|
2026-07-16 20:00 UTC |
Nvidia Cosmos 3 Edge boosts robot vision
According to @CNBC, Nvidia launched Cosmos 3 Edge, a real-time world model for robots and vision agents in Japan, enabling on-device perception and navigation. |
|
2026-07-06 20:22 UTC |
MIRA World Model Delivers 20 FPS Multiplayer
According to emollick, MIRA runs a four-player learned world at 20 FPS from keypresses, trained on 10k hours with bots, per General Intuition. |
|
2026-04-27 20:59 UTC |
KAI humanoid debuts 36-DOF hand, 18k sensors
According to AINewsOfficial, KAI revealed a 115-DOF humanoid with a 36-DOF hand, 18,000 tactile sensors at 0.1N precision, and a World Model brain. |
|
2026-01-29 09:23 UTC |
Latest Analysis: Project Genie Unveils Advanced World Model for Real-Time Playable Worlds
According to Demis Hassabis, Project Genie has been launched as an experimental prototype showcasing what is described as the world's most advanced world model, enabling users to generate entire playable worlds in real time from a simple text prompt. This technology, currently available to Ultra subscribers in the US, highlights potential breakthroughs in generative AI for interactive environments. As reported by Demis Hassabis on Twitter, Project Genie demonstrates significant business opportunities for gaming, simulation, and virtual world creation, leveraging advanced AI models for rapid content generation. |
|
2025-10-09 07:40 UTC |
Genie 3 AI Wins TIME Best Inventions 2025: Advanced World Model Generates Interactive Environments from Text and Image Prompts
According to @GoogleDeepMind, Genie 3 has been recognized by TIME as one of the Best Inventions of 2025, marking a significant advancement in AI-driven content generation. Genie 3 is a state-of-the-art world model capable of creating interactive, playable environments from simple text or image prompts, enabling new opportunities in gaming, virtual reality, and simulation-based training. The AI technology demonstrates practical applications for game developers and businesses seeking to streamline the creation of immersive digital experiences, potentially reducing development costs and accelerating innovation in the interactive media sector (source: @GoogleDeepMind, Oct 9, 2025; TIME). |