# Reframe: Observation-Aligned Multi-Teacher On-Policy Distillation for Agentic Visual Reasoning

**Authors:** Haozhi Yuan, Handong Li, Yifan Zhang, Qianshan Wei  
**Lab:** Nanjing University · **Published:** 9 Nov 2026 · **Upvotes:** 37  
**Links:** [arXiv](https://arxiv.org/abs/2611.12886) · [PDF](https://arxiv.org/pdf/2611.12886) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2611.12886.html)

## Key numbers

- **14**: Extensive experiments over 14 image and video benchmarks demonstrate that Reframe outperforms both Mix-RL and vanilla…

## Abstract

Agentic reinforcement learning (RL) and multi-teacher on-policy distillation (MOPD) offer complementary routes to post-training. For visual agents, however, tool use presents two related challenges. Models may overlook useful tool-returned observations, while poorly grounded actions can limit the evidence available for subsequent distillation. We introduce Reframe, a post-training framework that treats the observation state, rather than only the output distribution, as an object of alignment. First, Ground-Drop-GRPO (GD-GRPO) trains spatial and temporal specialists with evidence-aware rewards and stochastic removal of the original visual input, encouraging them to acquire grounded evidence and use the returned observations. Second, Observation-Gated-MOPD (OG-MOPD) replaces deficient tool actions with canonical ones mid-rollout and resumes distillation from the repaired observation. The two stages complement each other, as learning to rely on tool-returned observations can help the policy benefit from observation repair. Neither stage requires a task-specific SFT cold start. Extensive experiments over 14 image and video benchmarks demonstrate that Reframe outperforms both Mix-RL and vanilla MOPD. Ablation results show that GD-GRPO increases observation dependence and tool gain compared with outcome-based GRPO.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.