# openreasoner/openr

OpenR: An Open Source Framework for Advanced Reasoning with Large Language Models

Repository: https://github.com/openreasoner/openr
Canonical: https://ross.abutalabs.com/products/openr
Homepage: https://openreasoner.github.io/
Language: Python
License: MIT
License Family: permissive
Last push: 2025-01-17T19:08:11+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 2, release rhythm 35, longevity 49
- inputs: {"age_days": 691, "days_push": 593, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1853, forks 132 (observed 2026-08-28T04:05:44.563381+00:00)

## What it is
OpenR is an open-source Python framework that integrates search, reinforcement learning, and process supervision to improve chain-of-thought reasoning in large language models. It provides data pipelines, PRM training, MCTS-style search strategies, and test-time scaling experiments, accompanied by released datasets and models.

## Use cases
- train a process reward model for LLM reasoning steps
- run MCTS search over reasoning chains with an LLM
- fine-tune LLMs with online reinforcement learning for math reasoning
- generate reasoning step data from outcome labels automatically
- study test-time scaling laws for LLM inference
- improve model accuracy on math benchmarks like MATH

## When to choose
- you want an open-source reproduction of o1-style reasoning with RL and process supervision
- you need to train or evaluate process reward models
- you want to experiment with search strategies like MCTS at inference time
- you need reasoning datasets and pretrained PRMs to bootstrap research

## When to avoid
- you just need a general LLM inference or serving framework
- your task is simple prompting without multi-step reasoning
- you lack GPU resources for RL training and large-scale search
- you need a production chatbot rather than a research framework

## Facets
- artifact type: framework
- maturity: active
- function: llm-training, machine-learning, reinforcement-learning, rag, benchmarking, data-generation
- domain: large-language-models, machine-learning, reinforcement-learning, artificial-intelligence
- platform: python, cross-platform
- tags: process-supervision, prm, mcts, test-time-compute, reasoning, math-reasoning, reinforcement-learning, search-strategies, natural-language-processing, gpu, linux

## Member repositories
- openreasoner/openr (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:44.563381+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:16:50.809218+00:00, confidence not recorded.
  - readme: https://github.com/openreasoner/openr (fetched 2026-08-28T04:05:44.563381+00:00, sha 7e55970c38de)
  - homepage: https://openreasoner.github.io/ (fetched 2026-08-29T10:56:03.331743+00:00, sha 46d0a176cfc1)
- Data as of 2026-08-30T08:39:29.467469+00:00.