Ecosystem | NeMo Gym

NeMo Gym is one of several specialized libraries in the NVIDIA NeMo platform. Alongside NeMo RL for post-training, NeMo Gym defines the environments agents act in and composes them with agent harnesses so the same tasks and verification can drive both evaluation and training.

For what NeMo Gym does and when to use it, see the Overview and What NeMo Gym Provides. For how environments are implemented as composable servers, see Architecture.

NeMo Gym is also integrated with the broader agentic ecosystem. We would love your contribution! Open a PR to add an integration, or file an issue to share what would be valuable for you.

NeMo Gym Within NVIDIA NeMo

NVIDIA NeMo is a GPU-accelerated platform for training generative AI models and optimizing AI agents. It includes specialized libraries that span the model lifecycle end to end:

NeMo Curator: Prepares, filters, and deduplicates training data at scale.
NeMo Gym: Defines environments — tasks, verifier, tools, and state — and composes them with agents for evaluation and training (this library).
NeMo RL: Runs scalable post-training (GRPO, DPO, and SFT) against those environments.

NeMo Gym’s role: NeMo Gym is where a task becomes an interactive environment. It standardizes the environment surface, verification, and per-task state, then composes those with an agent harness to produce rollouts and reward signals for evaluation and training.

Environment Libraries

Seamlessly combine environments and benchmarks from other libraries alongside NeMo Gym environments, with access to over 1,000 community-contributed environments across integrated libraries.

Aviary
Harbor
OpenEnv
Reasoning Gym
Verifiers

Training Framework Libraries

Use environments for SFT and RL training. If you’re interested in integrating another training framework, see the Training Framework Integration Guide.

NeMo RL
Unsloth
VeRL

Agent Harnesses

Agent harnesses are available out of the box for evaluation and training. Some examples:

OpenHands - software engineering agent harness
Mini SWE Agent - software engineering agent harness
LangGraph - agent patterns built with LangGraph (reflection, orchestration, parallel thinking)

Depending on your workflow, you may also find these NeMo libraries useful across data preparation, evaluation and training, and deployment.

Data

Library	Purpose
NeMo Curator	Scalable data preprocessing and curation
NeMo Data Designer	Generate synthetic training data from scratch or seed examples
NeMo Safe Synthesizer	Generate privacy-safe synthetic copies of sensitive datasets
NeMo Anonymizer	Detect and replace sensitive data

Evaluation & Training

Library	Purpose
NeMo Gym	Evaluate and improve models and agents using environments (this project)
NeMo RL	Scalable post-training with GRPO, DPO, and SFT
NeMo Megatron-Bridge	Pretraining and fine-tuning with Megatron-Core
NeMo AutoModel	PyTorch native training for Hugging Face models

Deployment

Library	Purpose
NeMo Agent Toolkit	Connect and optimize teams of AI agents
NeMo Guardrails	Enforce safety and policy rules
NeMo Retriever	Document extraction and retrieval for RAG pipelines