Pinned Tweet
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 “what if?” before it acts.
More details: ifm.ai/pan/
We’re excited to collaborate with the NSF–Simons AI Institute for Cosmic Origins (@CosmicAI_Inst) at UT Austin on AI for Open Science, centered on a shared interest in AI for Open Science. 🔭
Together, we’ll advance foundation models to accelerate astronomical discovery. CosmicAI will apply IFM’s models to astronomy challenges and complex datasets, giving feedback on emerging AI capabilities in real-world science.
We’ll openly share the resulting models, built on the open foundation of K2 Horizon, and lessons learned with the scientific and open-source AI communities.
Here’s to AI as an engine for scientific discovery! ✨
Open weights are the bare minimum of open.
PAN 2 ships with the weights, the inference code, the training code and an Apache 2.0 license. Few open world models come with all four.
Inference code lets you run it. Training code lets you make it your own. Apache 2.0 lets you keep it.
huggingface.co/IFM/PAN2
Inference code available via:
• vLLM: github.com/vllm-project/vllm…
• Hugging Face Diffusers: huggingface.co/docs/diffuser…
• FastVideo: haoailab.com/FastVideo/cookb…
7
7
890
Several world model releases in a few months, and still not sure what a world model is for?
Here's one answer: it lets you ask "what if I try this?" and see the answer.
An LLM can tell you how to climb a route. It can't show you whether you'll make the reach.
PAN 2 does both, from one photo of the wall. 🧗
It’s made possible by two reasoners working together:
• Conceptual reasoner: plans the climb in words. Which holds to use, and which hand moves next on this route.
• Perceptual reasoner: shows visually on the wall the details the plan doesn't spell out: hips, reach, weight shift, timing. They decide whether the climb succeeds or fails.
👇 Concept demo. All climbing footage is generated by PAN 2.
Try PAN 2: github.com/ifm-ai/PAN2
Institute of Foundation Models retweeted
K2-Horizon: A Frontier Fleet of Open Source Language Models
Six fully open models from 0.9B to 375B-A23B MoE, with MoVA attention and Uno diffusion adapters, built for reasoning and agentic work. Full development history, data, and code released.
Looking for a reference manual for building a frontier LLM? We wrote one 👇
The K2 Horizon technical report is now on arXiv: 134 pages on how we built six models, from architecture to inference. It may be the most detailed LLM training reference you can read today.
• Architecture: we bring two new attention designs, MoVA and softplus attention gating, to strengthen reasoning. With them, our 36B model nearly matches a 32B dense one on one-seventh of the active parameters.
• Pre-training: we trained four models, from 3.7B to 36B, on the same 22T tokens. Their normalized loss curves collapse onto nearly the same line, so a recipe that works on a small model carries over when scaling up.
• Data: 60+ pages on the part most reports skip. How we generated ~10T synthetic tokens, and why we chose the data mix on predicted scores at scale, which reversed the small-run ranking.
• Post-training: we document our full production-scale run for the 375B flagship.
• Inference: Uno gives an autoregressive model diffusion speed with one small adapter. Up to 2.2× faster decoding, lossless by construction, no draft model.
This is the last piece of the openness puzzle. K2 Horizon already shipped with 380 checkpoints, 13 TB of pre-training data, 58 training runs and 17 code repositories. The report explains the decisions behind them.
More details: arxiv.org/abs/2611.21003
PAN 2 is now in Hugging Face Diffusers 0.43.0.
Text-to-world and image-to-world workflows for the 29B latent diffusion model, ready to run.
Thank you @huggingface Diffusers Team and @RisingSayak for the great collaboration.
github.com/huggingface/diffu…
1
10
1,114
A world model is only useful if it predicts what happens in the real world.
PAN 2 is built for faithful simulation. Agents plan by simulating ahead, so every step must be right, or errors compound.
• Multi-step plans: 🥇 #1 of all models tested, completing 34.0% of full plans, the most of any model
• Single actions: close second to MiniMax-H3 on six dimensions of physical intelligence (84.48% vs 87.20%)
• Blind human ratings: close second to MiniMax-H3 (Elo 1595 vs 1620)
Open weights, training code and inference code under Apache 2.0: huggingface.co/IFM/PAN2
Single actions: we ran an internal evaluation on six dimensions of physical intelligence.
PAN 2 scores 84.48% overall, second to MiniMax-H3 at 87.20%, and ranks in the top three on every dimension.
1
3
461
PAN 2 is the best open model in physics and robotics.
🥇 #1 open model on Physics-IQ Verified
🥇 #1 open models in our RBench evaluation
Open weights, training code and inference code under Apache 2.0: huggingface.co/IFM/PAN2
Thank you @haoailab for day-0 support of PAN 2 in FastVideo. With FastVideo, PAN 2 simulates 20 seconds in about 2 minutes on 8×B300, so developers can run it fast from the first day.
Try it now: huggingface.co/IFM/PAN2
FastVideo team is excited to announce day-0 support for @IFM_AI PAN 2! 🚀
PAN 2 is a foundation world model for physical intelligence, enabling simulative reasoning, where an agent imagines the outcomes of its candidate actions before choosing among them.
Simulate 20 seconds in 2 minutes on 8×B300. 10 seconds in 44 s. 📷
FastVideo: github.com/hao-ai-lab/FastVi…
PAN 2 Recipe: haoailab.com/FastVideo/cookb…
Congrats to the PAN 2 team at IFM!
@mbzuai
Huge thanks to @vllm_project for day-0 support of PAN 2.
Serve it on your own machine today:
huggingface.co/IFM/PAN2
🎉 Great work from the PAN 2 team on an open-weight world model built for real-world simulation, grounded in time, physics and space to predict how scenes evolve as actions unfold.
vLLM-Omni brings Day-0 support for PAN 2 text-to-video and image-to-video generation, making it available through an open inference stack.
PAN 2 is grounded in space.
Move the camera, and the world it shows doesn't change. Objects keep their identity and state, even when covered or out of view.
An agent can't plan in a world that rearranges itself whenever it looks away.
Below: a child stacks two blocks, then turns to look at the cat, which wanders around on its own. When the child looks back, the blocks are still stacked.





