What Nous Research does
Nous Research is one of the world's most active open-source AI labs. Founded in 2023, it became a community-led research collective that ships its own fine-tunes and post-trains of Llama, Mistral, and other open-weight base models — the Hermes series — which together have been downloaded more than 55 million times and now power 120+ mobile apps and products. In 2025, Nous shipped Hermes 4, a steerable instruction-tuned model family with minimal content restrictions, and launched the Hermes Agent, a long-running personal agent that grows context with the user. The lab is also pioneering decentralized training through Psyche, an open infrastructure built on Solana that pools compute from independent operators to train large models collaboratively.
Who it's for
Nous serves developers, indie AI builders, and open-source enthusiasts who want models with full transparency, strong steerability, and weight-level access — not API-only black boxes. Its decentralized training stack also targets researchers and infrastructure teams interested in collaborative, censorship-resistant compute markets.
Pricing
Nous models are free to download from Hugging Face. The Psyche network and Hermes Agent are open-source; commercial support and enterprise tuning are available through partnerships. There is no hosted SaaS pricing today.
Team & funding
Nous was founded by Jeffrey Quesnelle (CEO), Karan Malhotra (Head of Behavior), researcher Teknium, and Shivani Mitra. The lab raised $65M total in 2025 — including a $50M Series A led by Paradigm at a roughly $1B token valuation, with participation from Vipul Reddy (CEO of Together AI), Distributed Global, North Island Ventures, Delphi Digital, and Solana co-founder Raj Gokal. Funding fuels the Psyche decentralized training network, foundational compute infrastructure, and the next generation of Hermes models.