# Visual-OPSD: Cross-Modal On-Policy Self-Distillation for Efficient Unified Multimodal Reasoning

arXiv: https://arxiv.org/abs/2606.18974 · HTML: https://mlsanity.com/paper/2606.18974/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 29 of its 30 references that are in the map.
- First posted on arXiv 17 Jun 2026.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Knowledge distillation / [Multimodal distillation](https://mlsanity.com/distill/multimodal-distillation/)

## Builds on (most cited first, 29 of 29)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency](https://mlsanity.com/paper/2508.18265/) (2025)
- [ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning](https://mlsanity.com/paper/2203.10244/) (2022)
- [Chameleon: Mixed-Modal Early-Fusion Foundation Models](https://mlsanity.com/paper/2405.09818/) (2024)
- [Emerging Properties in Unified Multimodal Pretraining](https://mlsanity.com/paper/2505.14683/) (2025)
- [Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs](https://mlsanity.com/paper/2406.16860/) (2024)
- [Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling](https://mlsanity.com/paper/2501.17811/) (2025)
- [Emu3: Next-Token Prediction is All You Need](https://mlsanity.com/paper/2409.18869/) (2024)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs](https://mlsanity.com/paper/2401.06209/) (2024)
- [Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.01939/) (2025)
- [BLINK: Multimodal Large Language Models Can See but Not Perceive](https://mlsanity.com/paper/2404.12390/) (2024)
- [Born Again Neural Networks](https://mlsanity.com/paper/1805.04770/) (2018)
- [Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model](https://mlsanity.com/paper/2408.11039/) (2024)
- [V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs](https://mlsanity.com/paper/2312.14135/) (2023)
- [Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models](https://mlsanity.com/paper/2601.18734/) (2026)
- [Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation](https://mlsanity.com/paper/2410.13848/) (2024)
- [Teaching Small Language Models to Reason](https://mlsanity.com/paper/2212.08410/) (2022)
- [Visual Sketchpad: Sketching as a Visual Chain of Thought for Multimodal Language Models](https://mlsanity.com/paper/2406.09403/) (2024)
- [SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models](https://mlsanity.com/paper/2412.07755/) (2024)
- [Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation](https://mlsanity.com/paper/2605.18740/) (2026)
- [Visual Chain of Thought: Bridging Logical Gaps with Multimodal Infillings](https://mlsanity.com/paper/2305.02317/) (2023)
- [ThinkMorph: Emergent Properties in Multimodal Interleaved Chain-of-Thought Reasoning](https://mlsanity.com/paper/2510.27492/) (2025)
- [Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perception](https://mlsanity.com/paper/2602.11858/) (2026)
- [VSP: Assessing the dual challenges of perception and reasoning in spatial planning tasks for VLMs](https://mlsanity.com/paper/2407.01863/) (2024)
- [OPSDL: On-Policy Self-Distillation for Long-Context Language Models](https://mlsanity.com/paper/2604.17535/) (2026)

## Built on by (most cited first, 17 of 17)

- [OPD-V: Visual On-Policy Self-Distillation with Modality Balance](https://mlsanity.com/paper/2608.05131/) (2026)
- [VAD: Attributing Visual Evidence for Target Reconstruction in Multimodal On-Policy Distillation](https://mlsanity.com/paper/2607.28590/) (2026)
- [Where to Look Matters: On-Policy Self-Distillation for Long-Video Understanding](https://mlsanity.com/paper/2608.25356/) (2026)
- [OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution](https://mlsanity.com/paper/2609.06490/) (2026)
- [OPD-Aha: From Linguistic Momentum to Visual Reflection in Multimodal On-Policy Distillation](https://mlsanity.com/paper/2609.16459/) (2026)
- [Buried in Textual Debt: Context Pruning with Visual Evidence Preservation for MLLM Agents](https://mlsanity.com/paper/2608.22963/) (2026)
- [Aphanta: Diagnosing Task-Aligned Image-Edited Intermediates for Multimodal Reasoning](https://mlsanity.com/paper/2608.26993/) (2026)
- [Uni-LaDiR: Latent Diffusion Unifies Multimodal Reasoning](https://mlsanity.com/paper/2609.19878/) (2026)
- [SCOPD: Sparse-Context On-Policy Self-Distillation for Efficient Vision-Language Models](https://mlsanity.com/paper/2609.34044/) (2026)
- [Distilling Visual Reasoning into Text Space](https://mlsanity.com/paper/2609.34408/) (2026)
- [On-Policy Visual Evidence Distillation](https://mlsanity.com/paper/2609.36838/) (2026)
- [VISTA: Internalizing Collective Visual Experience via On-Policy Distillation for Active Multimodal Agents](https://mlsanity.com/paper/2609.38086/) (2026)
- [Omni-Diffusion-Distill: Few-Step Distillation of Unified Multimodal Diffusion Large Language Models](https://mlsanity.com/paper/2610.10990/) (2026)
- [SpatialOPSD: Self-Distilling Spatial Intelligence from Verified Coding Agent Traces](https://mlsanity.com/paper/2610.11366/) (2026)
- [Distilling Routed 3D Privilege for Spatial Reasoning in Vision-Language Models](https://mlsanity.com/paper/2610.12355/) (2026)
- [OWLET: On-Policy Weighted Learning via Evidence Transfer for Video Large Language Model Post-Training](https://mlsanity.com/paper/2610.13272/) (2026)
- [When MLLMs miss what humans read: diagnosing and learning to read squint text](https://mlsanity.com/paper/2610.13465/) (2026)

## Abstract (arXiv)

Unified multimodal models (UMMs) interleave generated ''visual thoughts'' (VTs) with text reasoning to improve spatial tasks. This incurs roughly an order-of-magnitude inference cost from multi-step diffusion. We find this cost yields limited direct benefit. On ThinkMorph, removing or noising VTs barely changes accuracy across nine benchmarks. Once rendered, attention concentrates on the VT regardless of content. Yet a KL diagnostic shows that conditioning on a privileged VT trace shifts the model's completion distribution. This suggests the generation pathway encodes useful reasoning beyond the rendered pixels. Motivated by this gap, we propose Visual On-Policy Self-Distillation(Visual-OPSD). Teacher and student share identical weights but differ in context: the teacher sees privileged VTs while the student sees only the question. Token-level JSD distillation on on-policy student trajectories transfers the teacher's reasoning to a text-only student. Across nine benchmarks, Visual-OPSD improves over its generative teacher by +3.40pp with 14.3× speedup (10.0s vs. 142.8s per sample) and outperforms same-scale VLMs by +63.83pp on VSP. A Gaussian-noise control (+0.40pp vs. +10.28pp for real VTs) and 58.4% closure of the KL gap confirm that gains come from the semantic content of the generation pathway.