Loop the Loopies!

Hugging Face Daily Papers Papers

Summary

Loopie introduces looped Mixture-of-Experts Transformers that outperform vanilla transformers under the same compute budget, achieving gold-medal performance at the 2025 IMO and IPhO without tools.

We present Loopie, the most powerful looped Transformer to date. The Loopie series consists of two Mixture-of-Experts (MoE) models: a 20B-parameter model with 2B active parameters and a 6Bparameter model with 0.6B active parameters. Looped Transformers have long faced a challenge: given an N-fold increase in pre-training compute, increasing the parameter count by a factor of N usually outperforms looping a model N times. Loopie addresses this challenge. Extensive ablation studies, including comparisons with a vanilla 30B-A3B model, show that Loopie substantially outperforms vanilla Transformer baselines trained with the same compute budget. Our novel post-training pipeline equips Loopie with strong reasoning abilities. At the 2025 IMO and IPhO, Loopie achieves gold-medal performance without tools.
Original Article
View Cached Full Text

Cached at: 07/20/26, 09:40 AM

Paper page - Loop the Loopies!

Source: https://huggingface.co/papers/2607.16051

Abstract

WepresentLoopie,themostpowerfulloopedTransformertodate.TheLoopieseriesconsistsoftwoMixture-of-Experts(MoE)models:a20B-parametermodelwith2Bactiveparametersanda6Bparametermodelwith0.6Bactiveparameters.LoopedTransformershavelongfacedachallenge:givenanN-foldincreaseinpre-trainingcompute,increasingtheparametercountbyafactorofNusuallyoutperformsloopingamodelNtimes.Loopieaddressesthischallenge.Extensiveablationstudies,includingcomparisonswithavanilla30B-A3Bmodel,showthatLoopiesubstantiallyoutperformsvanillaTransformerbaselinestrainedwiththesamecomputebudget.Ournovelpost-trainingpipelineequipsLoopiewithstrongreasoningabilities.Atthe2025IMOandIPhO,Loopieachievesgold-medalperformancewithouttools.

View arXiv pageView PDFAdd to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.16051 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.16051 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.16051 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Loop the Loopies!

arXiv cs.CL

Loopie is a new looped Transformer model that achieves gold-medal performance at the 2025 IMO and IPhO without external tools, using a novel post-training pipeline. It outperforms vanilla Transformers trained with the same compute budget.

SMELT: Scaling Laws for Compute-Matched MoE Looped Transformers

Hugging Face Daily Papers

SMELT is a method that loops middle layers in Mixture-of-Experts Transformers to improve training efficiency and downstream performance while matching compute, parameter, and cache budgets, leading to faster loss reduction and practical gains.

@askalphaxiv: Another cool research on Looped Transformers They ask the question: "Can we loop a frozen, off-the-shelf checkpoint dir…

X AI KOLs Timeline

This research introduces a technique to loop frozen, off-the-shelf transformer checkpoints at inference time by using damped Runge-Kutta substeps, treating transformer layers as Euler steps in a residual ODE. This allows extra latent compute without fine-tuning, architecture changes, or new weights, showing gains on knowledge tasks like MMLU-Pro, GPQA, and ARC.