Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
[NeurIPS 2025 Spotlight] LLM post-training suite — featuring ReasonFlux, ReasonFlux-PRM, and ReasonFlux-Coder.
| Date | Stars |
|---|---|
| 2026-07-31 | 541 |
| 2026-08-02 | 542 |
| 2026-08-06 | 542 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# 🧠 ReasonFlux Series ### *Advanced Open-Source LLM Post-Training Suite* **Princeton University** \& **PKU** \& **UIUC** \& **University of Chicago** \& **ByteDance Seed** **🎯 Mission**: Building next-generation reasoning capabilities through innovative LLM post-training algorithms focusing on **data selection**, **reinforcement learning**, and **inference scaling**. ## Contents of Repository - [Updates](#updates) - [Model Family Guide](#model-family-guide) - [ReasonFlux-PRM **(NeurIPS 2025)**](./ReasonFlux_PRM/README.md) - [ReasonFlux-Coder **(NeurIPS 2025 Spotlight)**](https://github.com/Gen-Verse/CURE) - [ReasonFlux](./ReasonFlux/README.md) - [Preliminary Work on Thought Template **(NeurIPS 2024 Spotlight)**](#preliminary-work-on-thought-template) - [Performance Hightlights](#performance-highlights) - [Citation](#citation) ## 🚀 What Makes ReasonFlux Series Special? ### 1. Trajectory-Aware Process Reward Models for Long-CoT Reasoning (ReasonFlux-PRM, NeurIPS 2025) Trajectory-aware reward models that provide dense supervision for both offline data selection and online policy optimization in long-CoT reasoning. <p align="center"> <img src="./ReasonFlux_PRM/img/intro_res.png" width=100%> </p> ### 2. Co-Evolved RL for LLM Coder and Unit Tester (ReasonFlux-Coder, NeurIPS 2025 Spotlight) Innovative approach where coders and unit testers evolve together through reinforcement learning, creating more robust coding capabilities. <p align="center"> <img src="./ReasonFlux_Coder/figures/overviewplot.png" width=100%> </p> ### 3. Long-CoT Reasoning with Thought Templates (ReasonFlux-Zero/F1) Revolutionary hierarchical reasoning framework that uses thought templates to guide complex problem-solving, achieving SOTA performance with higher efficiency. <p align="center"> <img src="./figs/comparison.png" width=100%> </p> ## Preliminary Work on Thought Template Our ReasonFlux-Zero/F1 models are built upon insights from our preliminary work on thought templates—specifically, [Buffer of Thoughts (NeurIPS 2024 Spotlight)](https://openreview.net/forum?id=ANO1i9JPtb) and [SuperCorrect (ICLR 2025)](https://openreview.net/forum?id=PyjZO7oSw2). These works introduce high-level, efficient intermediate reasoning patterns that guide and structure the thinking process of large language models. ## Updates - [2025/6/23] 🎉 We introduce [**ReasonFlux-PRM**](https://arxiv.org/abs/2506.18896), a family of trajectory-aware process reward models (PRMs) for long CoT reasoning in LLMs. ReasonFlux-PRM is able to support **both offline and online reward supervision**, by selecting high-quality training data for model distillation, providing dense process-level rewards for policy optimization during reinforcement learning, and enabling reward-guided test-time scaling. Our trained PRMs including [ReasonFlux-PRM-7B](https://huggingface.co/Gen-Verse/ReasonFlux-PRM-7B) and [ReasonFlux-PRM-1.5B](https://huggingface.co/Gen-Verse/ReasonFlux-PRM-1.5B) are now available on [HuggingFace-GenX](https://huggingface.co/Gen-Verse). We also release a 7B advanced thinking and reasoning model [ReasonFlux-PRM-Qwen-2.5-7B](https://huggingface.co/Gen-Verse/ReasonFlux-PRM-Qwen-2.5-7B) supervised via our PRM. - [2025/6/04] 🎉 We release our [**Co-Evolving RL**](https://github.com/Gen-Verse/CURE) optimized coding LLMs, [ReasonFlux-Coder-7B](https://huggingface.co/Gen-Verse/ReasonFlux-Coder-7B) and [ReasonFlux-Coder-14B](https://huggingface.co/Gen-Verse/ReasonFlux-Coder-14B), which outperform similarly sized Qwen Coders and DeepSeek Coders, and naturally fit into common test-time scaling and agentic coding pipelines. We also release our Long-CoT model [ReasonFlux-Coder-4B](https://huggingface.co/Gen-Verse/ReasonFlux-Coder-4B), outperforming Qwen3-4B while achieving 64.8% efficiency in unit test generation. - [2025/3/24] 🎉We release [ReasonFlux-F1-32B](https://huggingface.co/Gen-Verse/ReasonFlux-F1), [ReasonFlux-F1-14B](h
Excerpt of 10,387 characters
Read on GitHub
Ling Yang
44
Zhaochen Yu · National University of Singapore · Singapore
7
Jiaru (Rubin) Zou · Stanford University
2
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:049614398a551f34, topic:reinforcement-learning
matched fp:049614398a551f34, topic:code-generation