Paper 2511.16043
Agent0: Unleashing Self-Evolving Agents from Zero Data via Tool-Integrated Reasoning
- Published
- Nov 2025
- Research lab
- Independent
- Citations
- 54
- GitHub
- 1.2K stars
01 In brief
Summary
Agent0 is a fully autonomous framework that evolves LLM agents from zero external data by combining tool-integrated reasoning with multi-step co-evolution.
It initializes two agents from the same base LLM: a curriculum agent trained via GRPO to generate frontier tasks, rewarded by the executor's uncertainty (self-consistency), tool-use frequency, and a repetition penalty; and an executor agent trained via ADPO (ambiguity-dynamic policy optimization) on filtered challenging tasks using majority-vote pseudo-labels and a sandboxed code interpreter.
The tool integration creates a virtuous cycle: improved executor capability pressures the curriculum agent to generate harder, tool-dependent tasks.
Experiments on Qwen3-4B-Base and Qwen3-8B-Base show Agent0 outperforms baselines like R-Zero, Absolute Zero, SPIRAL, and Socratic-Zero, improving mathematical reasoning by 18% and general reasoning by 24% on Qwen3-8B-Base.
Ablations confirm the importance of curriculum training, tool reward, repetition penalty, ADPO, and multi-turn reasoning.
Task difficulty and tool calls increase across iterations, demonstrating progressive co-evolution.
02 From the paper
Abstract
Large Language Model (LLM) Agents, often trained with Reinforcement Learning (RL), are constrained by a dependency on human-curated data, limiting scalability and tethering AI to human knowledge. Existing self-evolution frameworks offer an alternative but are typically restricted by the model's inherent capabilities and single-round interactions, hindering the development of complex curricula involving tool use or dynamic reasoning. We introduce Agent0, a fully autonomous framework that evolves high-performing agents without external data through multi-step co-evolution and seamless tool integration. Agent0 establishes a symbiotic competition between two agents initialized from the same base LLM: a curriculum agent that proposes increasingly challenging frontier tasks, and an executor agent that learns to solve them. We integrate external tools to enhance the executor's problem-solving capacity; this improvement, in turn, pressures the curriculum agent to construct more complex, tool-aware tasks. Through this iterative process, Agent0 establishes a self-reinforcing cycle that continuously produces high-quality curricula. Empirically, Agent0 substantially boosts reasoning capabilities, improving the Qwen3-8B-Base model by 18% on mathematical reasoning and 24% on general reasoning benchmarks. Code is available at https://github.com/aiming-lab/Agent0.