Organization
Nous Research
Model developer · Research lab · Developer platform · United States (city not stated; the store privacy policy gives a contact address in Austin, Texas)
Nous Research is an AI research company that trains and releases open-weight language models, most prominently the Hermes series of post-trained models, and builds open-source agent and training software such as Hermes Agent and the Atropos reinforcement-learning environment framework.1289
Key facts
- Ownership
- Privately held
- Legal form
- Corporation established under the laws of Delaware, per the Nous Portal terms of service.3
- Headquarters
- United States (city not stated; the store privacy policy gives a contact address in Austin, Texas)4
- Other documented locations
- New York, New York5
- Founded
- Unknown
- Sectors
- Open models, Agents, Research
What is and isn’t public
Nous Research publishes Hermes model weights on Hugging Face. These are post-trained from other organizations' base models; the model cards declare, for example, Apache 2.0 for the Seed-OSS-based Hermes 4.3 36B and a Llama license ("llama3") for the Llama-based Hermes 4 70B. It also publishes evaluation samples and open-source tools such as Atropos and Hermes Agent. The Hermes 4 technical report describes the post-training data but does not link a release of it. This catalog assesses openness per release.11121398
Products
Hosted model products
Labeled Closed product/API — this describes how a documented product is delivered (hosted access). It is not an open artifact and says nothing about other releases by the same organization.
| Product | Kind | Documented delivery and access | Reviewed |
|---|---|---|---|
| Nous Portal (external site: portal.nousresearch.com) | Hosted model API | Account and credits-based subscription service that provides model inference (including the Nous Inference API), hosted tools, and cloud hosting for Hermes Agent; its catalog includes Nous models and many third-party models.6311 | Sep 29, 2026 |
Other products
| Product | Kind | Documented delivery and access | Reviewed |
|---|---|---|---|
| Hermes Agent (external site: hermes-agent.nousresearch.com) | Developer tool | Open-source (MIT) autonomous agent that runs from a terminal install, desktop apps, or cloud hosting through Nous Portal, connects to messaging platforms, and can use Nous Portal or another model provider.78 | Sep 29, 2026 |
Related open artifacts
Model families
- HermesModel family
Model releases
- Hermes 4 70BModel releaseAvailability: PublicModel-disclosure tier (USASI rubric v0.1): Open-weight
- Hermes 4.3 36BModel releaseAvailability: PublicModel-disclosure tier (USASI rubric v0.1): Open-weight
Notable documented facts
- Nous Research says Hermes 4.3 36B was its first Hermes model trained over the internet on Psyche, a decentralized training network, and that it also released a centrally trained version for comparison.1011
- For Hermes 4, Nous Research published the evaluation samples generated during benchmarking on Hugging Face alongside the models, as described in the Hermes 4 technical report.13
U.S. eligibility
Eligible · basis: U.S. headquarters
The Nous Portal terms of service identify the operator as Nous Research, Inc., a Delaware corporation. The only postal address found in Nous Research's own documents is 600 Congress Avenue, Austin, Texas, listed as a contact address for Nous Research, Inc. in its store privacy policy; the careers page refers to a New York office and lists other roles as remote. No official page reviewed names a headquarters city, so only the country is recorded; no non-U.S. office was found, and the U.S. headquarters basis rests on these documents.345
Sources
This listing is not an endorsement or a federal approval.