European Open Source AI Index
DatabaseNewsGuidesAboutContribute

Guru

by LLM360

Guru, a model which claims to improve upon the state-of-the-art.
Text
Latest
https://huggingface.co/LLM360/guru-32B
Guru-32B
Qwen2.5-32B
CC-BY-NC-4.0
LLM360, an initiative to fully open-source LLMs.
https://www.llm360.ai/index.html
May 2025
Availability
Base Model Data
Pretraining data not specified or documented.
End User Model Data
Dataset published on HuggingFace.
https://huggingface.co/datasets/LLM360/guru-RL-92k
Base Model Weights
Model weights made available on HuggingFace.
https://huggingface.co/Qwen/Qwen2.5-32B
End User Model Weights
Model weights made available on HuggingFace.
https://huggingface.co/LLM360/guru-32B
Training Code
Base model repository provides sparse source code and some examples for SFT. End model repository contains training script
https://github.com/QwenLMhttps://github.com/LLM360/Reasoning360/tree/main/scripts/train
Documentation
Code Documentation
Both repositories are fairly well-documented.
https://github.com/QwenLMhttps://github.com/LLM360/Reasoning360
Hardware Architecture
Hardware architecture described for end model.
https://arxiv.org/pdf/2506.14965https://github.com/LLM360/Reasoning360/blob/main/scripts/train/example_multinode_rl_qwen32b_base.sh
Preprint
Preprints published on arXiv.
https://arxiv.org/abs/2505.09388https://arxiv.org/pdf/2506.14965
Paper
Peer-reviewed paper submitted to NeurIPS.
https://openreview.net/forum?id=xUBgfvyip3
Modelcard
Model card primarily contains usage instructions.
https://huggingface.co/LLM360/guru-32B
Datasheet
Detailed datasheet provided on HuggingFace.
https://huggingface.co/datasets/LLM360/guru-RL-92k
Access
Licenses
CC-BY-NC-4.0
https://huggingface.co/LLM360/guru-32B
Is this information not up to date?
Contribute here ->
Supported by the Centre for Language Studies and the Dutch Research Council. Website design & development © 2024 by BSTN. This version of the index generated 09 April 2026, website content last updated 11 March 2026.