Excited to finally share PAN 2! ๐
Pretrained on millions of hours of egocentric, robotic, and human-motion data, PAN 2 is a step toward foundation world models that can understand and predict the physical world.
Introducing PAN 2: an open foundation world model for physical intelligence.
A language model predicts the next word. PAN 2 predicts the next world: what happens after each action, grounded in time, physics and space.
Towards AI that simulates โwhat if?โ before it acts.
More details: ifm.ai/pan/
Zeyu Feng(Ted) retweeted
๐ Introducing ๐๐ช๐ฎ๐ข๐ฅ๐ข๐๐ซ๐ข๐ฎ๐ฆ ๐๐๐๐ฌ๐จ๐ง๐๐ซ๐ฌ (๐๐ช๐) !
Feedforward models and weight-tied models behave very differently on hard reasoning generalization.
EqR pushes this difference to the extreme by learning ๐ญ๐๐ฌ๐ค-๐๐จ๐ง๐๐ข๐ญ๐ข๐จ๐ง๐๐ ๐ง๐๐ฎ๐ซ๐๐ฅ ๐๐ญ๐ญ๐ซ๐๐๐ญ๐จ๐ซ๐ฌ .
โข Sudoku-Extreme: 99.8%
โข Maze: 93%
#ICML2026
Zeyu Feng(Ted) retweeted
Agents can now live like humans in a virtual city built with Unreal Engine: exercising outdoors, strolling through parks, or even playing chess with friends!
๐ Check out SimWorld: an open-ended simulator for LLM agents to act, perceive, and plan in richly embodied, endlessly diverse virtual worlds.
SimWorld now supports plug-and-play integration with arbitrary UE environments. Deploy agents in your own scenes and build fully customizable simulations for autonomous driving, collaborative games, or social experiments!
SimWorld is a strong testbed for evaluating embodied agents, as it provides:
- โ๏ธPhysical + social simulation: realistic physics (collision/friction/inertia) and social dynamics like traffic rules/flows.
- ๐ค LLM/VLM-friendly interfaces: a Gym-style API with multimodal observations and grounded language actions.
- ๐ค Long-horizon tasks + reasoning: diverse physical/social tasks for systematic training and evaluation.
1/
#LLM #Agent #UnrealEngine #simulation
Zeyu Feng(Ted) retweeted
Pretraining has scaling laws to guide compute allocation. But for RL on LLMs, we lack a practical guide on how to spend compute wisely.
We show the optimal compute allocation in LLM RL scales predictably.
โ Key takeaways below
Zeyu Feng(Ted) retweeted
๐ Excited to announce the release of PAN, a general world model Iโve been working on for years.
PAN can simulate physical, agentic, and nested worlds โ generating infinite interactive experiences to train and evaluate AI agents.
Check out demo: ifm.mbzuai.ac.ae/pan/
๐
๐ Introducing long-horizon, interactable PAN World Model that lets you step into coherent worlds and evolve them through language-guided actions.
A world model is far more than visuals โ it understands world dynamics, enabling an agent to imagine scenarios, anticipate outcomes, and act with strategy.
Check out our website & paper ๐
Website: ifm.mbzuai.ac.ae/pan/
Paper: arxiv.org/abs/2511.09057
Zeyu Feng(Ted) retweeted
๐ฅReally excited to see the release of PAN world model, a project I had been working over the past years.
PAN is a general world model capable of simulating physical, agentic, and nested worlds, synthesizing infinite interactive experiences for training AI agents.
Building on top of pretrained LLMs and video diffusion models, PAN connects language, perception, action, and latent thoughts, for long-horizon simulation and reasoning.
PAN shows overwhelming performance gains over JEPA-2, Cosmos-2, and other prior models. More in the thread๐ ... 1/
Zeyu Feng(Ted) retweeted
๐ Introducing PAN, our latest general world model.
๐ก Compared to traditional video generation models like Sora 2, PAN simulates worlds you can interact with, over long horizons, with natural-language actions.
Zeyu Feng(Ted) retweeted
Super excited to introduce Pandora, a generative video World Model interactively controllable by language.
#Sora and #GPT4 are both powerful. How about fusing them in a single model? ๐ฅ
Pandora gives a preview:๐ญ
> Build a General World Model (GWM) super efficiently by integrating pretrained autoregressive LLM and diffusion Video Model, aligning them in the representation space.
> Let the LLM control the VM on-the-fly. Instruction tuning maximizes the controllability.
> Autoregressive LLM empowers VM to generate indefinitely long videos: Starting with a VM for 2-second videos, Pandora extends it for 8-second videos. Would #Sora+#GPT4 under Pandora produce hours-long videos? ๐ฝ๏ธ
> World Model is beyond just video generation. Itโs sensory-level information processing + concept-level reasoning and reflection. Pandora bridges both, with the concept- / language-level backbone (LLM) managing & steering the sensory-level VM functionalities. ๐๏ธ๐ง
Check out for a bunch of interesting results:
world-model.ai
๐ฅIntroducing Pandora ๐ ๐ช
a World Model that generates videos of world states with real-time language control ๐ฅ๐น๏ธ
Simulate the world across domains in an _interactive_ way!
check out more world-model.ai
I am excited to share the latest work from our group:
๐ฃ๐ฎ๐ป๐ฑ๐ผ๐ฟ๐ฎ๐.
It is a world model that predicts future states with real-time steerable control using natural language. It is a step towards General World Model.
๐ฅIntroducing Pandora ๐ ๐ช
a World Model that generates videos of world states with real-time language control ๐ฅ๐น๏ธ
Simulate the world across domains in an _interactive_ way!
check out more world-model.ai
3
211
Zeyu Feng(Ted) retweeted
๐๐๐Today, we're thrilled to introduce ๐๐๐ง๐๐จ๐ซ๐: a ๐ฌ๐ค๐ง๐ก๐ ๐ข๐ค๐๐๐ก๐ that generates videos๐ฅ with ๐ง๐๐๐ก-๐ฉ๐๐ข๐ control with ๐๐ง๐๐-๐ฉ๐๐ญ๐ฉ actions๐ฌ!!!
Check it out more world-model.ai !
@MaitrixOrg
Zeyu Feng(Ted) retweeted
๐ฅ We are excited to announce ๐ฃ๐ฎ๐ป๐ฑ๐ผ๐ฟ๐ฎ, a world model with natural language actions and video states.
๐ It is a step towards a General World Model that:
1. Simulates world states by generating videos across any domains
2. Allows any-time control with free-text actions
Zeyu Feng(Ted) retweeted
๐ฅIntroducing Pandora ๐ ๐ช
a World Model that generates videos of world states with real-time language control ๐ฅ๐น๏ธ
Simulate the world across domains in an _interactive_ way!
check out more world-model.ai









