Of interest to maths researchers and Putnam-style problem solvers; anyone else has little reason to download it.
Kenning's take
What it does
Nomos 1 is a version of Qwen3-30B-A3B-Thinking tuned by Nous Research, with Hillclimb AI, for maths problem-solving and proof writing. Used with the Nomos reasoning framework, which was open-sourced at the same time, it scored 87 out of 120 on the 2025 Putnam exam, against 24 for the base model under the same conditions. The weights are free on Hugging Face under Apache 2.0.
- Built on Qwen3-30B-A3B-Thinking-2507
- Maths problem-solving and natural-language proofs
- Scored 87/120 on Putnam 2025 with its reasoning framework
- Open-source Nomos reasoning framework
- Apache 2.0 weights, runs with Transformers, vLLM or SGLang
Who it suits. Researchers and developers working on AI for mathematics who can run a 30B model.
Key facts
- Category
- AI models
- Made by
- Nous Research
- Launched
- 2025
- How to get it
- free download from Hugging Face for self-hosting
- Works on
- self-hosted
- In the UK
- Yes
- Free plan
- free open weights
FAQ
What is Nomos 1?
An open maths reasoning model from Nous Research, built on Qwen3-30B-A3B-Thinking and tuned for problem-solving and proofs.
How well does Nomos 1 do on the Putnam?
Nous Research reports 87 out of 120 on Putnam 2025 when used with its Nomos reasoning framework.
Is Nomos 1 free?
Yes. The weights are on Hugging Face under the Apache 2.0 licence.
Facts from the maker's own pages and the sources listed. Checked 2026-10-11.
Similar tools
ACE-Step 1.5
Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.
AlphaGenome Atlas
Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.
Amazon Nova 2
Amazon's second family of Nova AI models, covering reasoning, speech conversation and combined text-and-image generation.
Cisco Antares
Cisco's small open-weight models that find where known security vulnerabilities sit in a codebase, running locally.