Nous Research
AI research organization focused on open-source model training, fine-tuning, and agent infrastructure. Based in the US.
Key projects
- Hermes models — fine-tuned LLMs optimized for function calling and tool use (the “Hermes” family). Used as the base for tool-calling training and evaluation.
- Hermes Agent — open-source, self-improving AI agent framework with multi-platform gateway, skills system, and RL training infrastructure.
- Nous Portal — hosted inference API for Nous models and third-party models.
- agentskills.io — open standard and hub for agent skills (procedural memory).
- Atropos/Tinker — RL training framework for tool-calling models, integrated into Hermes Agent as
tinker-atropossubmodule.
Research contributions
- Tool-call parser implementations for multiple model families (Hermes, DeepSeek, Qwen, Llama, Mistral, GLM, Kimi, Longcat)
- Benchmark evaluation environments (TBLite, Terminal-Bench 2.0, YC-Bench)
- Trajectory compression for synthetic training data generation
- Collaboration with OpenThoughts, Collinear AI, Snorkel AI, Bespoke Labs on benchmark datasets
Links
- Website: nousresearch.com
- Discord: discord.gg/NousResearch
- GitHub: github.com/NousResearch