On-policy Distillation (OPD) supervises a student model on trajectories sampled from its own policy by minimizing the divergence between the output distributions of the teacher and student at each token position, thereby providing dense token-level supervision. Although existing OPD methods have demonstrated strong performance in improving the reasoning ability of student models, their objectives fundamentally rely on token-level distribution matching.
Contrastive On-Policy Distillation (COPD) is a framework that improves on-policy distillation by using a frozen teacher to evaluate student states under two contrasting prompts—one encouraging low reasoning effort and one encouraging high effort. The difference in log‑probabilities between these prompts provides a token‑level advantage signal that guides the student toward more concise and efficient reasoning strategies. Experiments on nine multimodal benchmarks show that COPD reduces reasoning length while maintaining task performance, and the contrastive approach can also be applied to on‑policy self‑distillation, allowing a model to compress its own reasoning without an external teacher.
By Jiacheng Ruan, Jun Tang, Wenzhen Yuan, Ting Liu, Shuai Bai, Dayiheng Liu, Zhibo Yang, Yuzhuo Fu
arXiv:2606. 05718v1 Announce Type: cross Abstract: On-policy distillation (OPD) improves reasoning by training a student on trajectories sampled from its own policy under supervision from a teacher.
By Kanghui Tian, Siyuan Liu, Ziang Yan, Sheng Xia, Shuai Dong, Yi Wang
arXiv:2607. 02592v1 Announce Type: cross Abstract: On-policy distillation (OPD) has recently emerged as an effective post-training paradigm by providing supervision on student-generated trajectories.
By Qixiang Yin, Huanjin Yao, Yuchen Cai, Jianghao Chen, Ziyi Wang, Min Yang, Fei Su, Zhicheng Zhao
Privileged on-policy distillation improves multimodal reasoning by allowing a teacher to evaluate student trajectories using rich, training-only visual evidence. Both models score these trajectories w...
arXiv:2609.36838v1 Announce Type: cross
Abstract: Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from a strong teache...
By Shaohang Wei, Feifan Song, Guangyue Peng, Wenhao Yu, Wei Li, Wen Luo, Yang Xu, Yufan Shen, Luke Mao, Yang Du, Asher Qin, Houfeng Wang
arXiv:2608. 01263v1 Announce Type: new Abstract: On-policy distillation (OPD) samples trajectories from the current student policy and minimizes token-level divergence between student and teacher next-token distributions at prefixes along those trajectories.
By Leyan Xue, Feng Xiong, Mingjun Ma, Changqing Zhang
OPD‑Aha is a privileged on‑policy distillation method that improves multimodal reasoning by reconstructing the distillation target from the teacher’s isolated visual preference instead of relying on fragile teacher‑student discrepancies. It suppresses continuations that contradict the image, encouraging students to interrupt flawed reasoning with reflection tokens such as "wait" and "actually." This approach leads to consistent improvements across fine‑grained perception and complex multimodal reasoning benchmarks.
By Chenhao Qiu, Dawei Li, Yechao Zhang, Lei Gong, Zhen Tan
Video-OPSD introduces a post‑training framework for Video Large Language Models that leverages privileged visual evidence to enhance on‑policy self‑distillation. The method constructs a self‑teacher conditioned only on annotated evidence frames, while the student processes the full video, allowing the teacher to provide more focused supervision. Additionally, an evidence‑guided token optimization scheme weights distillation based on each token’s reliance on privileged evidence, improving perceptually grounded reasoning. Experiments demonstrate consistent gains over standard OPSD and comparable performance to GRPO with less training time.
By Ziyue Wang, Shiqi Huang, Weiwen Xu, Bihan Wen, Xudong Jiang
CA-OPD is a confidence‑aware on‑policy distillation framework that improves structured visual prediction by using teacher confidence to selectively correct unreliable student transitions and gradually transfer rollout control to the student. The method aligns supervision with intervention decisions, providing direct cross‑entropy loss for corrected tokens and full predictive distribution for retained tokens. In a multi‑teacher setting for GUI grounding and OCR, CA‑OPD significantly outperforms the Qwen3.5‑0.8B baseline, achieving large gains on benchmarks such as ScreenSpot‑Pro and OCRBench‑v2 English.
By Menghao Li, Linjie Mu, Yin Wang, Haotian Hu, Yannian Gu, Lujiayi Xue, Fanyi Wang
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.
arXiv:2609.25891v1 Announce Type: new
Abstract: Multimodal large language models (MLLMs) often struggle with fine-grained visual perception when processing complete images, as critical evidence may o...
By Zihan Chen, Hengguang Zhou, Yuan Kang, Yiming Zhang, Wenhui Fang, Zenghui Ding, Yining Sun, Cho-Jui Hsieh