HackerPapers

arXiv:2501.19393cs.CL

s1: Simple test-time scaling

Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, Tatsunori Hashimoto

Abstract

Test-time scaling is a promising new approach to language modeling that uses extra test-time compute to improve performance. Recently, OpenAI's o1 model showed this capability but did not publicly share its methodology, leading to many replication efforts. We seek the simplest approach to achieve test-time scaling and strong reasoning performance. First, we curate a small dataset s1K of 1,000 questions paired with reasoning traces relying on three criteria we validate through ablations: difficulty, diversity, and quality. Second, we develop budget forcing to control test-time compute by forcefully terminating the model's thinking process or lengthening it by appending "Wait" multiple times to the model's generation when it tries to end. This can lead the model to double-check its answer, often fixing incorrect reasoning steps. After supervised finetuning the Qwen2.5-32B-Instruct language model on s1K and equipping it with budget forcing, our model s1-32B exceeds o1-preview on competition math questions by up to 27% (MATH and AIME24). Further, scaling s1-32B with budget forcing allows extrapolating beyond its performance without test-time intervention: from 50% to 57% on AIME24. Our model, data, and code are open-source at https://github.com/simplescaling/s1

arXiv comments
46 pages (9 main), 10 figures, 15 tables
Journal reference
DOI
Subjects
cs.CLComputation and Language cs.AIArtificial Intelligence cs.LGMachine Learning

On Hacker News

Linked by 12 identifiable commenters in 12 comments across 11 threads, from Feb 2025 to Jul 2026. Found in 3 HN submissions from Feb 2025 to Jun 2025. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments11 threads

  1. Controlling Reasoning Effort in LLMs 1 comment

    • > My favorite trick for controlling the reasoning level is the hack First described here I think, specifically for Qwen models: https://arxiv.org/abs/2501.19393, although it seems unclear how much it actually improves the final response unless you create a new fine-tuned model…

      embedding-shapeView on HNParent context

  2. The One-Step Trap (In AI Research) 1 comment

    • This is the same reasoning behind why Yann Lecun thought test-time scaling would not work for LLMs: compounding error. Instead, the more tokens LLMs use, the better their performance on many tasks. LLMs can self-correct, evidenced by the power of getting models to question…

      mxwsnView on HNParent context

  3. TinyLoRA – Learning to Reason in 13 Parameters 1 comment

    • >One theory is that the knowledge required to solve the task is already stored in the parameters of the model, and only the style has to change for task success >In particular, learning to generate longer outputs may be possible in few parameters Reminded me of…

      kgeistView on HNParent context

  4. Chain of Recursive Thoughts: Make AI think harder by making it argue with itself 1 comment

    • a paper with a similar idea on scaling test time reasoning, this is sorta how all the thinking models work under the hood. https://arxiv.org/abs/2501.19393

      mangomanView on HNParent context

  5. Tencent's 'Hunyuan-T1'–The First Mamba-Powered Ultra-Large Model 1 comment

    • I don’t think you have an accurate understanding of how LLMs work. https://arxiv.org/abs/2501.19393 These tokens DO extend the thinking time. We are talking about causal autoregressive language models, and so these tokens can be used to guide the generation.

      computerexView on HNParent context

  6. QwQ-32B: Embracing the Power of Reinforcement Learning 1 comment

    • If that's an issue, there's a workaround using structure generation to force it to output a </thiking> token after some threshold and force it to write the final answer. It's a method used to control thinking token generation showcased in this paper…

      lelagView on HNParent context

  7. LIMO: Less Is More for Reasoning 1 comment

    • Yes, the authors explicitly highlighted those two points in the abstract, in terms of them being the elicitation threshold for complex reasoning, namely, an extremely complete pre-trained foundation model, and a set of extremely high quality examples post-training. To your…

      armcatView on HNParent context

  8. The LLM Curve of Impact on Software Engineers 1 comment
  9. Researchers created an open rival to OpenAI's o1 'reasoning' model for under $50 1 comment

    • nope you have the article right. "if you copy an existing model, you can get a pretty comparable model!" Really, [what the underlying paper](https://arxiv.org/abs/2501.19393) shows is how much you can improve existing models with several of the techniques released by openai…

      wrenkyView on HNParent context

  10. S1: A $6 R1 competitor? 2 comments

    • In case you’re not sure what S1 is, here is the original paper: https://arxiv.org/html/2501.19393v1

      bberenbergView on HNParent context

    • Agreed. Here are three things that I find surreal about the s1 paper. (1) The abstract changed how I thought about this domain (advanced reasoning models). The only other paper that did that for me was the "Memory Resource Management in VMware ESX Server". And that paper got…

      ozguneView on HNParent context

  11. Chat is a bad UI pattern for development tools 1 comment

    • The file explains it a bit, but my blogpost https://xeiaso.net/notes/2025/s1-simple-test-time-scaling/ could probably be better explained. I'll write out more but just for you I'll summarize what I'm gonna end up writing up. AI models fundamentally work on the basis of "given…

      xenaView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.