Guan Wang

🚀Introducing Hierarchical Reasoning Model🧠🤖 Inspired by brain's hierarchical processing, HRM delivers unprecedented reasoning power on complex tasks like ARC-AGI and expert-level Sudoku using just 1k examples, no pretraining or CoT! Unlock next AI breakthrough with neuroscience. 🌟 📄Paper: arxiv.org/abs/2506.21734 💻Code: github.com/sapientinc/HRM

Hierarchical reasoning works well on large language models!🎉

🔥It’s official-Sapient HRM Discord Community is now live！ This is a place to discuss, connect, and collaborate as we shape HRM’s future together. We will be sharing our latest work, releases, and tips, as well as hosting Q&A sessions💬💬 Hop on this journey with us as we push the boundaries of what HRM and AGI at large can achieve！🙌 ➡️Join us on Discord here discord.gg/sapient

Thanks to @arcprize for reproducing and verifying the results! ARC-AGI-1: public 41% pass@2 - semi private 32% pass@2 ARC-AGI-2: public 4% pass@2 - semi private 2% pass@2 Due to differences in testing environments, a certain amount of variance in results is acceptable. According to tests run on our infrastructure, the open-source version of HRM on our GitHub can achieve a score of 5.4% pass@2 on the ARC-AGI-2. We welcome everyone to run it on your own infra and share your scores~ This is our first submission to the leaderboard, and it's a good starting point. We appreciate everyone for your support and feedback on HRM, both before and after our appearance on the ARC leaderboard. All of this encourages and motivates us to improve. The hierarchical architecture is designed to resolve premature convergence in long-horizon tasks, like master-level Sudoku that takes hours for humans to solve. See the comparison with a simple recurrent Transformer. Such a long chain might not be essential for ARC problems, and we only used a high-low ratio of 1/2. Larger ratios are often needed for optimal performance for Sudoku problems. In the case of ARC-AGI, the success of HRM is a testament to the model's ability to exhibit fluid intelligence - that is, its capability to infer and apply abstract rules from independent and flat examples. We are glad it was discovered in a recent blog post that the outer loop and data augmentation are essential for this ability, and we especially thank @fchollet @GregKamradt @k_schuerholt for pointing this out. Finally, we are accelerating the iteration of the HRM model and continuously pushing its limits, with good progress so far. At the same time, we believe the hierarchical architecture is highly effective in many scenarios. Moving forward, we will make further targeted updates to the architecture and validate it on more applications. We will also release an FAQ to address the key questions raised by the community. 🧠 Stay tuned!

❤️ Thanks for highlighting our HRM paper! Apologies for any confusion—we're working on clarifying this thoroughly. Stay tuned for updates!

@taihongtran Search just finds stuff. HRM learns patterns and reasons. We’re testing if it can work as a search after training - stay tuned.

@makingAGI What is the differences compare to neural search algorithms?

🚀Introducing Hierarchical Reasoning Model🧠🤖 Inspired by brain's hierarchical processing, HRM delivers unprecedented reasoning power on complex tasks like ARC-AGI and expert-level Sudoku using just 1k examples, no pretraining or CoT! Unlock next AI breakthrough with neuroscience. 🌟 📄Paper: arxiv.org/abs/2506.21734 💻Code: github.com/sapientinc/HRM

@evilmathkid We have clarified it here: github.com/sapientinc/HRM…

@makingAGI @makingAGI The writing is a little unclear, could you please clarify which dataset are the ARC results on? Public train, public eval, private eval? Can't find you on the leaderboard on kaggle

@ADarmouni No. Eval data stayed separate. Details here: github.com/sapientinc/HRM…

@makingAGI Something I need clarification: are the Arc-AGI results on data in the training set? You mention public eval in train data, so I suppose you trained on it But is it also the data you evaluate it on? Even if it overfits the training data, results are super cool!

@Homo_Immortalis Answers will not be coming soon, but we’re working on it!

@makingAGI This is exciting Guan, when do you think AI will be able to answer humanity’s unsolvable questions, like reversing aging, or understanding consciousness?

@NewAgeNihilism Thanks! Any questions, email us anytime: research@sapient.inc.

@makingAGI Great job, really good paper. I'm still wrapping my head around it

@codeslubber Fair point! We’re inspired by brains, not handmade expert-system hierarchies. We let the hierarchy learn and grow from data, not hand-made rules.

Will read the paper, but saying that by making the code hierarchical we are using neuroscience is kind of silly no? I am super interested in how these hierarchies manifest. One of the reasons Expert Systems failed was there never was any consensus on how to organize knowledge representation, and encoding everything in a bespoke fashion was not scalable.

@mikeboysen HRM is a way smaller, focused model: A) needs less data, B) answers niche questions without a giant LLM, C) runs cheap.

@makingAGI What does this mean for consumer cost?

@itsAlexGuerrero @grok LLMs training = auto-regressive paradigm. HRM training = RNN thinker with a built-in “stop” signal (ACT) so it knows when to stop. Different playbooks.

@makingAGI @grok how does this differ from the traditional way LLMs are trained? What does this mean for the future of AI?

@smjain Love it! 🎞️ Your TL;DR helps - sharing it.

@makingAGI wonderful..Here i created a small animation to summarize the concept as per what I understood

@notrajivpoddar Nope. LLMs are awesome but architecture-capped. We’re trying to push past those limits.

@makingAGI so can i use an llm to generate the samples and train it at test time and provide true general intelligence?

@mysticaltech @ylecun Tagging the legend himself - thanks for the shout-out!

@makingAGI This is properly revolutionary. @ylecun FYI, right down your alley.

@JonathanRoseD No plan yet. But code, checkpoints, and demo data are open - feel free to roll your own 🛠️

@makingAGI Will there be a GGUF model release?

Thanks for featuring us!😃

🌟Exactly my thoughts on the next-gen of AI reasoning. Leveraging insights from neuroscience, our Hierarchical Reasoning Model offers practical, efficient scaling of depth. Not every problem can be solved faster with more processors; sometimes, the key is adding depth.

Our co-founder William Chen is going to share more about the open-sourced Hierarchical Reasoning Model (HRM) at #FortuneAISingapore @FortuneMagazine tomorrow, under the panel theme "Beyond Human: AGI And The Future We’re Building"! We are excited about the practical path towards universally capable reasoning systems that rely on architectures, not scale, to reach real AGI. ⏰16:10-16:40 SGT, July 23, Mainstage

@ai_for_success Only ~2 GPU hours for pro Sudoku. 50~200 for ARC-AGI 😀

@makingAGI what was the total cost of training?? Also would be interested in full breakdown post.