← Back to Research
arxiv
LLM February 22, 2026 Impact: 8/10

Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach

By Jonas Geiping, Sean McLeish, Neel Jain, John Kirchenbauer, Siddharth Singh, Brian R. Bartoldson, Bhavya Kailkhura, Abhinav Bhatele, Tom Goldstein
We study a novel language model architecture that is capable of scaling test-time computation by implicitly reasoning in latent space. Our model works by iterating a recurrent block, thereby unrolling to arbitrary depth at test-time. This stands in contrast to mainstream reasoning models that scale up compute by producing more tokens. Unlike approaches based on chain-of-thought, our approach does not require any specialized training data, can work with small context windows, and can capture types of reasoning that are not easily represented in words. We scale a proof-of-concept model to 3.5 billion parameters and 800 billion tokens, showing the resulting model can improve its performance on reasoning benchmarks, sometimes dramatically, up to a computation load equivalent to 50 billion parameters.
test-time computelatent reasoningrecurrent depthreasoning modelsarchitecture
Analysis

Summary

This paper from researchers at the University of Maryland and Lawrence Livermore National Laboratory introduces a fundamentally different approach to test-time scaling. Rather than having models "think longer" by generating more chain-of-thought tokens (the approach used by o1, DeepSeek-R1, and similar systems), they propose a recurrent depth architecture where a single block is iterated repeatedly at inference time. The model can unroll to arbitrary depth, spending more compute without producing any additional output tokens. Their proof-of-concept reaches 3.5B parameters trained on 800B tokens.

Why It Matters

The dominant paradigm for reasoning models right now — generate more tokens, think out loud — has real limitations. It requires specialized training data with reasoning traces, burns through context window budget, and fundamentally can only capture reasoning patterns that can be expressed in natural language. This latent reasoning approach sidesteps all three constraints. There's no need for curated chain-of-thought datasets, context window is not consumed, and the model can potentially internalize reasoning patterns that are genuinely hard to verbalize. The result: on some reasoning benchmarks, the 3.5B model performs equivalently to a 50B parameter model when given sufficient compute budget at inference time.

Implications

If this approach scales well, it represents a third path for reasoning AI — alongside pure scale (bigger models) and explicit chain-of-thought (more tokens). Latent reasoning could be particularly valuable in constrained environments: embedded systems, real-time applications, or situations where verbose chain-of-thought output is undesirable. It also raises interesting questions about interpretability — when reasoning happens entirely in latent space, there's no scratchpad to inspect, which could complicate alignment and safety work.

Open Questions

The paper is an explicit proof-of-concept, and the authors are candid about this. Key unknowns include whether the approach scales to larger base models, whether it can match the raw benchmark performance of dedicated reasoning models like o1 or R1, and whether the latent representations developed during iterative unrolling are stable and predictable. The fact that they released model weights and code suggests strong community confidence — we'll likely see follow-up scaling studies from independent researchers soon.

Key Findings
1
Recurrent depth architecture enables test-time scaling without generating additional tokens
2
3.5B parameter model achieves performance equivalent to 50B parameter model with sufficient inference compute
3
No specialized reasoning training data required — works with standard language model pretraining
4
Latent reasoning captures patterns not easily expressed in natural language, unlike chain-of-thought
Impact Score
8/10
Expert Commentary
This is one of the more architecturally interesting papers of early 2025 — it challenges the assumption that better reasoning necessarily means more verbose reasoning, and opens a genuinely new direction for efficient inference scaling.
D
Dong Tran
AI Researcher. Deep-dive analysis of frontier AI research papers.