# verl-agent

> verl-agent is an extension of veRL, designed for training LLM/VLM agents via RL. verl-agent is also the official code for paper "Group-in-Group Policy Optimization for LLM Agent Training"

## Facts
- **Category:** [Agent Frameworks](https://agentsearchengine.app/category/agent-frameworks)
- **Type:** Framework
- **Pricing:** Open source
- **GitHub stars:** 2,194
- **Language:** Python
- **License:** Apache-2.0
- **Last repo activity:** 2 months ago (Jun 9, 2026)
- **Official site:** https://huggingface.co/papers/2505.10978
- **Source code:** https://github.com/langfengQ/verl-agent
- **Listing on Agent Search Engine:** https://agentsearchengine.app/agents/verl-agent
- **Data verified:** July 2026

## About verl-agent
verl-agent is an extension of veRL, specifically designed for training large language model (LLM) agents via reinforcement learning (RL). Unlike prior approaches that simply concatenate full interaction histories, verl-agent proposes step-independent multi-turn rollout mechanism, which allows for fully customizable per-step input structures, history management, and memory modules. This design makes verl-agent highly scalable for very long-horizon, multi-turn RL training (e.g., tasks in ALFWorld can require up to 50 steps to complete). verl-agent provides a diverse set of RL algorithms (including our new algorithm GiGPO) and a rich suite of agent environments, enabling the development of reasoning agents in both visual and text-based tasks.

_From the project's README._

## Install

```
pip3 install vllm==0.11.0
```

## Alternatives
verl-agent sits in agent frameworks. Related tools indexed here:
- [langchain](https://agentsearchengine.app/agents/langchain) — The agent engineering platform. · Open source · Framework · 144k★ · Python · MIT · active 7 days ago
- [ruflo](https://agentsearchengine.app/agents/ruflo) — The leading agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. · Open source · Framework · 68k★ · TypeScript · MIT · active 7 days ago
- [autogen](https://agentsearchengine.app/agents/autogen) — A programming framework for agentic AI · Open source · Framework · 60k★ · Python · CC-BY-4.0 · active 4 months ago
- [crewAI](https://agentsearchengine.app/agents/crewai) — Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks. · Open source · Framework · 57k★ · Python · MIT · active 7 days ago
- [pydantic-ai](https://agentsearchengine.app/agents/pydantic-ai) — AI Agent Framework, the Pydantic way · Open source · Framework · 19k★ · Python · MIT · active 7 days ago
- [eliza](https://agentsearchengine.app/agents/eliza) — Open source agentic operating system · Open source · Framework · 19k★ · TypeScript · MIT · active 7 days ago
- [SuperAGI](https://agentsearchengine.app/agents/superagi) — <> SuperAGI - A dev-first open source autonomous AI agent framework. Enabling developers to build, manage & run useful autonomous agents quickly and reliably. · Open source · Framework · 18k★ · Python · MIT · active 1 year ago
- [agent-framework](https://agentsearchengine.app/agents/agent-framework) — A framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and.NET. · Open source · Framework · 13k★ · Python · MIT · active 6 days ago

---
Source: [Agent Search Engine](https://agentsearchengine.app) — an independent, hand-curated index of AI agents, MCP servers, and frameworks. Rankings are never sold; sponsored placements are labeled and never numbered. Full directory: https://agentsearchengine.app/llms-full.txt