← Back to Research
arxiv
LLM February 22, 2026 Impact: 9/10

s1: Simple Test-Time Scaling

By Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, Tatsunori Hashimoto
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).
test-time scalingreasoningbudget forcingo1-replicationQwen2.5
Analysis

Summary

The s1 paper tackles one of the most tantalizing mysteries in modern AI: how does OpenAI's o1 model achieve its remarkable reasoning performance through test-time compute scaling? The Stanford/UW team's answer is surprisingly elegant — you don't need massive proprietary infrastructure. With just 1,000 carefully curated training examples and a deceptively simple technique called "budget forcing," they fine-tune Qwen2.5-32B into s1-32B, a model that outperforms o1-preview on competitive mathematics benchmarks by up to 27%. The core insight is that quality and diversity of reasoning traces matter far more than quantity.

Why It Matters

Budget forcing is the paper's most novel contribution and deserves special attention. By appending the word "Wait" to interrupt a model's premature conclusion, the researchers force it to continue deliberating — essentially compelling the model to double-check its work. This trivially-simple intervention reliably improves accuracy on hard problems. It reframes test-time scaling not as a complex architectural challenge but as a compute allocation problem: you can trade inference time for accuracy in a controllable, predictable way. This democratizes o1-style reasoning to anyone with access to open models.

Implications

The s1K dataset curation methodology is equally significant. The team's three-criteria filter — difficulty, diversity, and quality — yields a dataset 1,000x smaller than what one might expect yet achieves frontier-level results. This suggests we've been dramatically over-estimating the data requirements for reasoning fine-tuning. The ability to extrapolate beyond training performance (50% to 57% on AIME24) by simply scaling budget forcing at inference time also hints that capable reasoning may be latent in large pre-trained models, waiting to be unlocked rather than trained from scratch.

Broader Context

This paper lands at a critical moment when the AI community is actively trying to replicate and understand o1-style reasoning without OpenAI's blessing. s1 joins DeepSeek-R1 and ReasonFlux as open-source answers to the o1 question, each taking a distinct approach. Where DeepSeek-R1 uses pure RL and ReasonFlux uses hierarchical thought templates, s1 bets on minimal supervised data plus clever inference control. The convergence of these independent approaches on test-time scaling as a real phenomenon — not just an OpenAI artifact — is one of the most important validation signals in recent AI research.

Key Findings
1
Budget forcing — appending 'Wait' to extend model thinking — improves reasoning accuracy on hard math problems with zero additional training
2
Only 1,000 curated training examples (s1K) are sufficient to achieve test-time scaling, dramatically lower than expected
3
s1-32B exceeds o1-preview on MATH and AIME24 by up to 27%, using an open-source base model (Qwen2.5-32B)
Impact Score
9/10
Expert Commentary
s1 is arguably the cleanest demystification of o1-style reasoning yet published — the budget forcing technique alone is worth the read, and the 1K dataset result should make everyone rethink data scaling assumptions.
D
Dong Tran
AI Researcher. Deep-dive analysis of frontier AI research papers.