🛰️ Daily AI Frontier
‹ back to 2026-07-20

Loop the Loopies!

Research LLMs & Foundation Models

Merged summary

TL;DR - Loopie is a series of looped Mixture-of-Experts Transformers that reportedly outperforms standard Transformers under equal compute budgets. Its post-training pipeline also yields tool-free gold-medal performance on the 2025 IMO and IPhO.

  • Includes 20B/2B-active and 6B/0.6B-active MoE models.
  • Reuses Transformer layers through looping rather than simply scaling parameter count.
  • Ablations include comparisons against a vanilla 30B-A3B model.
  • Claims to overcome the historical compute-efficiency disadvantage of looped Transformers.

Sources (1)

Loop the Loopies!

arXiv cs.CL Zitian Gao, Yilong Chen, Yihao Xiao, Xinyu Yang, Ran Tao, Joey Zhou, Bryan Dai 2026-07-17 arXiv:2607.16051

TL;DR - Loopie is a series of looped Mixture-of-Experts Transformers that reportedly outperforms standard Transformers under equal compute budgets. Its post-training pipeline also yields tool-free gold-medal performance on the 2025 IMO and IPhO.

  • Includes 20B/2B-active and 6B/0.6B-active MoE models.
  • Reuses Transformer layers through looping rather than simply scaling parameter count.
  • Ablations include comparisons against a vanilla 30B-A3B model.
  • Claims to overcome the historical compute-efficiency disadvantage of looped Transformers.
item →