Daily Papers of 2026-06-26

  1. DanceOPD: On-Policy Generative Field Distillation 80 upvotes, #1 of 2026-06-26
  2. In-Context World Modeling for Robotic Control 61 upvotes, #2 of 2026-06-26
  3. OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning 54 upvotes, #3 of 2026-06-26
  4. Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation 49 upvotes, #4 of 2026-06-26
  5. The Verification Horizon: No Silver Bullet for Coding Agent Rewards 47 upvotes, #5 of 2026-06-26
  6. ViQ: Text-Aligned Visual Quantized Representations at Any Resolution 38 upvotes, #6 of 2026-06-26
  7. JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting 35 upvotes, #7 of 2026-06-26
  8. GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents 28 upvotes, #8 of 2026-06-26
  9. Fast LeWorldModel 27 upvotes, #9 of 2026-06-26
  10. Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments 18 upvotes, #10 of 2026-06-26
  11. Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It 18 upvotes, #10 of 2026-06-26
  12. Discretizing Reward Models 17 upvotes, #12 of 2026-06-26
  13. LISA: Likelihood Score Alignment for Visual-condition Controllable Generation 13 upvotes, #13 of 2026-06-26
  14. Information-Aware KV Cache Compression for Long Reasoning 11 upvotes, #14 of 2026-06-26
  15. Confidence-Aware Tool Orchestration for Robust Video Understanding 11 upvotes, #14 of 2026-06-26
  16. PhysiFormer: Learning to Simulate Mechanics in World Space 11 upvotes, #14 of 2026-06-26
  17. Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents 10 upvotes, #17 of 2026-06-26
  18. CoffeeBench: Benchmarking Long-Horizon LLM Agents in Heterogeneous Multi-Agent Economies 9 upvotes, #18 of 2026-06-26
  19. Hallucination in World Models is Predictable and Preventable 8 upvotes, #19 of 2026-06-26
  20. ABACUS: Adapting Unified Foundation Model for Bridging Image Count Understanding and Generation 5 upvotes, #20 of 2026-06-26
  21. COrigami: An AI Pipeline for Co-Designing Flat-Foldable Visually Recognisable Origami 5 upvotes, #20 of 2026-06-26
  22. OpenBioRQ: Unsolved Biomedical Research Questions for Agents 4 upvotes, #22 of 2026-06-26
  23. When Does Combining Language Models Help? A Co-Failure Ceiling on Routing, Voting, and Mixture-of-Agents Across 67 Frontier Models 4 upvotes, #22 of 2026-06-26
  24. How Post-Training Shapes Biological Reasoning Models 3 upvotes, #24 of 2026-06-26
  25. EO-WM: A Physically Informed World Model for Probabilistic Earth Observation Forecasting 3 upvotes, #24 of 2026-06-26

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.