# LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling

arXiv: https://arxiv.org/abs/2606.18023 · HTML: https://mlsanity.com/paper/2606.18023/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Efficient reasoning and agent inference, one of 12 directions in Efficient LLM inference and serving.
- Builds on 18 of its 19 references that are in the map.
- First posted on arXiv 16 Jun 2026.

## Placed in

- Efficient LLM inference and serving / [Efficient reasoning and agent inference](https://mlsanity.com/inference/efficient-reasoning-and-agent-inference/)

## Builds on (most cited first, 18 of 18)

- [Universal Transformers](https://mlsanity.com/paper/1807.03819/) (2018)
- [Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach](https://mlsanity.com/paper/2502.05171/) (2025)
- [Looped Transformers as Programmable Computers](https://mlsanity.com/paper/2301.13196/) (2023)
- [Looped Transformers are Better at Learning Learning Algorithms](https://mlsanity.com/paper/2311.12424/) (2023)
- [Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA](https://mlsanity.com/paper/2410.20672/) (2024)
- [Latent Chain-of-Thought? Decoding the Depth-Recurrent Transformer](https://mlsanity.com/paper/2507.02199/) (2025)
- [Parallel Loop Transformer for Efficient Test-Time Computation Scaling](https://mlsanity.com/paper/2510.24824/) (2025)
- [How Much Is One Recurrence Worth? Iso-Depth Scaling Laws for Looped Language Models](https://mlsanity.com/paper/2604.21106/) (2026)
- [Stabilizing Recurrent Dynamics for Test-Time Scalable Latent Reasoning in Looped Language Models](https://mlsanity.com/paper/2605.26733/) (2026)
- [From Code Foundation Models to Agents and Applications: A Comprehensive Survey and Practical Guide to Code Intelligence](https://mlsanity.com/paper/2511.18538/) (2025)
- [LoopUS: Recasting Pretrained LLMs into Looped Latent Refinement Models](https://mlsanity.com/paper/2605.11011/) (2026)
- [Memory-Efficient Looped Transformer: Decoupling Compute from Memory in Looped Language Models](https://mlsanity.com/paper/2605.07721/) (2026)
- [Two-Scale Latent Dynamics for Recurrent-Depth Transformers](https://mlsanity.com/paper/2509.23314/) (2025)
- [Solve the Loop: Attractor Models for Language and Reasoning](https://mlsanity.com/paper/2605.12466/) (2026)
- [Loop as a Bridge: Can Looped Transformers Truly Link Representation Space and Natural Language Outputs?](https://mlsanity.com/paper/2601.10242/) (2026)
- [LT2: Linear-Time Looped Transformers](https://mlsanity.com/paper/2605.20670/) (2026)
- [Hierarchical vs. Flat Iteration in Shared-Weight Transformers](https://mlsanity.com/paper/2604.14442/) (2026)
- [CART: Context-Anchored Recurrent Transformer -- A Parameter-Efficient Architecture with Learned Stability](https://mlsanity.com/paper/2606.01495/) (2026)

## Built on by (most cited first, 4 of 4)

- [Improving Test-Time Scaling with Adaptive Looped Transformers](https://mlsanity.com/paper/2609.35748/) (2026)
- [Penelope: Localized Latent Recurrence for Efficient Structured Reasoning](https://mlsanity.com/paper/2607.25915/) (2026)
- [WaveFront Decoding: Parallelized Self-Speculative Decoding for Looped Language Models](https://mlsanity.com/paper/2609.23033/) (2026)
- [Decoding Looped Transformers Better for (Almost) Free](https://mlsanity.com/paper/2610.02185/) (2026)

## Abstract (arXiv)

Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop Transformers (PLT) alleviate this cost through cross-loop position offsets (CLP) and shared-KV gated sliding-window attention, making loop count a practical design choice. We therefore study PLT loop-count selection through a gain--cost view: an extra loop may refine representations, but CLP also introduces a positional mismatch at each loop boundary. We instantiate this study by training LoopCoder-v2, a family of 7B PLT coders with different loop counts, from scratch on 18T tokens, followed by matched instruction tuning and evaluation. Empirically, the two-loop variant delivers broad gains over the non-looped baseline across code generation, code reasoning, agentic software engineering, and tool-use benchmarks, improving SWE-bench Verified from 43.0 to 64.4 points and Multi-SWE from 14.0 to 31.0 points. In contrast, variants with three or more loops regress, revealing a strongly non-monotonic loop-count effect. Our diagnostics show that loop 2 provides the main productive refinement, while later loops yield diminishing, oscillatory updates and reduced representational diversity. Because the CLP-induced mismatch remains roughly fixed as refinement gains shrink, the offset cost increasingly dominates. This gain--cost trade-off explains PLT's saturation at two loops and provides diagnostics for loop-count selection.