# Diffusion reasoning and test-time scaling

Area: [LLM reasoning](https://mlsanity.com/lr/index.md). HTML: https://mlsanity.com/lr/diffusion-reasoning-and-test-time-scaling/

## Key facts

- Diffusion reasoning and test-time scaling: 130 papers on arXiv since 2022 (data through 14 Oct 2026).
- 32 of its 130 mapped papers (25%) cite “d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning” (2025, arXiv 2504.12216).
- 24 of its 130 mapped papers (18%) cite “A General Framework for Inference-time Scaling and Steering of Diffusion Models” (2025, arXiv 2501.06848).
- 20 new papers 17 Sep–14 Oct.

## Most cited here

- [d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning](https://mlsanity.com/paper/2504.12216/) (2025), cited by 32 papers here
- [A General Framework for Inference-time Scaling and Steering of Diffusion Models](https://mlsanity.com/paper/2501.06848/) (2025), cited by 24 papers here
- [Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps](https://mlsanity.com/paper/2501.09732/) (2025), cited by 23 papers here
- [wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models](https://mlsanity.com/paper/2507.08838/) (2025), cited by 17 papers here
- [Remasking Discrete Diffusion Models with Inference-Time Scaling](https://mlsanity.com/paper/2503.00307/) (2025), cited by 16 papers here
- [Mercury: Ultra-Fast Language Models Based on Diffusion](https://mlsanity.com/paper/2506.17298/) (2025), cited by 16 papers here
- [Beyond Autoregression: Discrete Diffusion for Complex Reasoning and Planning](https://mlsanity.com/paper/2410.14157/) (2024), cited by 15 papers here
- [Reinforcing the Diffusion Chain of Lateral Thought with Diffusion Language Models](https://mlsanity.com/paper/2505.10446/) (2025), cited by 13 papers here
- [SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer](https://mlsanity.com/paper/2501.18427/) (2025), cited by 10 papers here
- [Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing](https://mlsanity.com/paper/2503.19385/) (2025), cited by 10 papers here

## New (17 Sep–14 Oct)

- [Are We Overthinking One-Step Generation?](https://mlsanity.com/paper/2610.20331/) (2026-10-14)
- [Imperfect Verifiers, Exact Targets: Inference-Time Scaling for Continuous Generation](https://mlsanity.com/paper/2610.15281/) (2026-10-11)
- [Anchors Enhance Controllability and Reasoning of Diffusion Language Models](https://mlsanity.com/paper/2610.15097/) (2026-10-10)
- [RAMP: Recurrent Adaptive Mixed Precision for Efficient Prefill in Recurrent-Depth Language Models](https://mlsanity.com/paper/2610.13041/) (2026-10-09)
- [Rethinking the Roles of Diffusion and Decoding: Predictive-State Flows for Parallel Language Generation](https://mlsanity.com/paper/2610.12898/) (2026-10-09)
- [DiffusionART: Efficient, Stable and Scalable Reinforcement of Diffusion Models](https://mlsanity.com/paper/2610.12572/) (2026-10-08)
- [Think Before You Paint: Recursive Latent Reasoning for Diffusion Models](https://mlsanity.com/paper/2610.09876/) (2026-10-07)
- [Preserving Mathematical Reasoning in Compressed Diffusion Language Models via Trajectory-Aware Low-Rank Approximation](https://mlsanity.com/paper/2610.03326/) (2026-10-02)
- [Refinement Buys Intelligibility, Search Buys Identity: What Test-Time Compute Buys in Masked-Diffusion TTS](https://mlsanity.com/paper/2610.03320/) (2026-10-02)
- [Architectural Sampling: Test-Time Scaling via Computational Diversity in Frozen Vision-Language Models](https://mlsanity.com/paper/2610.01687/) (2026-10-01)
- [Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models](https://mlsanity.com/paper/2610.00661/) (2026-09-30)
- [ExploreNet: Learning Where to Explore in Diffusion GRPO](https://mlsanity.com/paper/2609.38329/) (2026-09-29)
- [Rethinking Soft Tokens for Parallel Decoding in Diffusion Language Models](https://mlsanity.com/paper/2609.37391/) (2026-09-29)
- [Why Deterministic PRM Guidance Underperforms in Discrete Diffusion Reasoning](https://mlsanity.com/paper/2609.35472/) (2026-09-28)
- [ForkLeft: Entropy-First Rollouts for Prefix-Aligned Autoregressive-to-Diffusion Distillation](https://mlsanity.com/paper/2609.32448/) (2026-09-26)
- [ELF-REG: Scaling Continuous Diffusion Language Models to Reasoning Tasks](https://mlsanity.com/paper/2609.29102/) (2026-09-24)
- [Towards Efficient Reasoning: Learning Causal Shortcuts for Diffusion Language Models](https://mlsanity.com/paper/2609.28272/) (2026-09-23)
- [Leveraging Inference-Time Compute for Diffusion Models via Global Scheduling of Denoising Trajectories](https://mlsanity.com/paper/2609.22867/) (2026-09-19)
- [λ-Controlled GRPO: Turning Flow-Matching Ratio Instability into a Budgeted Resource](https://mlsanity.com/paper/2609.22041/) (2026-09-18)
- [CaLR: Causal Latent Revision for Robust Diffusion Reasoning](https://mlsanity.com/paper/2609.20981/) (2026-09-17)