# Co-Evolving Agents: Learning from Failures as Hard Negatives

arXiv: https://arxiv.org/abs/2511.22254 · HTML: https://mlsanity.com/paper/2511.22254/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 33 of its 43 references that are in the map.
- First posted on arXiv 27 Nov 2025.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 30 of 33)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents](https://mlsanity.com/paper/2207.01206/) (2022)
- [Self-Rewarding Language Models](https://mlsanity.com/paper/2401.10020/) (2024)
- [Contrastive Learning with Hard Negative Samples](https://mlsanity.com/paper/2010.04592/) (2020)
- [Scaling Relationship on Learning Mathematical Reasoning with Large Language Models](https://mlsanity.com/paper/2308.01825/) (2023)
- [LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders](https://mlsanity.com/paper/2404.05961/) (2024)
- [AgentTuning: Enabling Generalized Agent Abilities for LLMs](https://mlsanity.com/paper/2310.12823/) (2023)
- [ScienceWorld: Is your Agent Smarter than a 5th Grader?](https://mlsanity.com/paper/2203.07540/) (2022)
- [FireAct: Toward Language Agent Fine-tuning](https://mlsanity.com/paper/2310.05915/) (2023)
- [WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning](https://mlsanity.com/paper/2411.02337/) (2024)
- [InterCode: Standardizing and Benchmarking Interactive Coding with Execution Feedback](https://mlsanity.com/paper/2306.14898/) (2023)
- [Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents](https://mlsanity.com/paper/2403.02502/) (2024)
- [Advancing LLM Reasoning Generalists with Preference Trees](https://mlsanity.com/paper/2404.02078/) (2024)
- [Generalized Preference Optimization: A Unified Approach to Offline Alignment](https://mlsanity.com/paper/2402.05749/) (2024)
- [RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation](https://mlsanity.com/paper/2306.11706/) (2023)
- [Agent-FLAN: Designing Data and Methods of Effective Agent Tuning for Large Language Models](https://mlsanity.com/paper/2403.12881/) (2024)
- [DART-Math: Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving](https://mlsanity.com/paper/2407.13690/) (2024)
- [Learn-by-interact: A Data-Centric Framework for Self-Adaptive Agents in Realistic Environments](https://mlsanity.com/paper/2501.10893/) (2025)
- [Watch Every Step! LLM Agent Learning via Iterative Step-Level Process Refinement](https://mlsanity.com/paper/2406.11176/) (2024)
- [Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training](https://mlsanity.com/paper/2501.11425/) (2025)
- [AgentRefine: Enhancing Agent Generalization through Refinement Tuning](https://mlsanity.com/paper/2501.01702/) (2025)
- [DynaSaur: Large Language Agents Beyond Predefined Actions](https://mlsanity.com/paper/2411.01747/) (2024)
- [3D-Properties: Identifying Challenges in DPO and Charting a Path Forward](https://mlsanity.com/paper/2406.07327/) (2024)

## Built on by (most cited first, 8 of 8)

- [Verified Critical Step Optimization for LLM Agents](https://mlsanity.com/paper/2602.03412/) (2026)
- [Self-Evolving World Models for LLM Agent Planning](https://mlsanity.com/paper/2606.30639/) (2026)
- [SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories](https://mlsanity.com/paper/2606.01311/) (2026)
- [On-Policy Self-Evolution via Failure Trajectories for Agentic Safety Alignment](https://mlsanity.com/paper/2605.11882/) (2026)
- [Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries](https://mlsanity.com/paper/2607.10113/) (2026)
- [Interactive Critique-Revision Training for Reliable Structured LLM Generation](https://mlsanity.com/paper/2605.08327/) (2026)
- [CWM: Contrastive World Models for Action Feasibility Learning in Embodied Agent Pipelines](https://mlsanity.com/paper/2602.22452/) (2026)
- [DeLoop: Repeated Actions as Self-Supervision for Long-Horizon Agents](https://mlsanity.com/paper/2610.19376/) (2026)

## Abstract (arXiv)

Self-evolving agents improve their performance on long-horizon tasks by learning from their own interactions with an environment. A common approach uses the resulting failed trajectories as negatives for preference training. However, collecting an agent's own failures does not ensure that they provide informative supervision for further improvement. Obvious failures may be easy to reject without learning to identify errors in more plausible attempts. These plausible but incorrect trajectories can serve as hard negatives, helping agents learn to distinguish successful behavior from failures with less obvious errors. To enable agents to generate and learn from such informative negatives, we propose a co-evolving framework that couples a target agent's self-improvement with a failure agent trained to generate hard negatives. The failure agent learns exclusively from both agents' failed trajectories to generate higher-reward failures, which the target agent uses as negatives for preference training. As the target policy improves, its failures provide new training data for the failure agent. The updated failure agent, in turn, provides new negatives for target training, allowing negative generation to adapt alongside the target policy. Across online shopping, scientific reasoning, and interactive SQL querying, our framework improves average task reward by 5.7 over the baseline. These results demonstrate the value of learning to generate and use informative failures for agent self-improvement. We will make our code publicly available.