# OpenMOSS/MOSS

An open-source tool-augmented conversational language model from Fudan University

Repository: https://github.com/OpenMOSS/MOSS
Canonical: https://ross.abutalabs.com/products/moss
Homepage: https://txsun1997.github.io/blogs/moss.html
Language: Python
License: Apache-2.0
License Family: permissive
Topics: chatgpt, deep-learning, dialogue-systems, large-language-models, natural-language-processing, text-generation
Last push: 2026-05-27T13:16:14+00:00

## Health v2 (maintenance only)
Score: 68/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 84, release rhythm 35, longevity 88
- inputs: {"age_days": 1236, "days_push": 98, "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 12230, forks 1127 (observed 2026-08-28T04:10:52.089773+00:00)

## What it is
MOSS is an open-source tool-augmented conversational large language model from Fudan University, released with base models, SFT models, plugin-augmented variants, quantized checkpoints, and SFT dialogue datasets. The repository provides inference code for local deployment and fine-tuning scripts for adapting the models to custom dialogue data.

## Use cases
- run a chatgpt-like conversational model locally
- deploy an open-source LLM on my own GPU
- fine-tune a dialogue model on my own conversation data
- use a language model with plugins like search and text-to-image
- run a quantized LLM on 12GB of VRAM
- download open-source Chinese-English instruction-tuned model weights
- generate code and answers with a self-hosted chatbot

## When to choose
- you want a fully open, self-hosted ChatGPT-style model with weights and training data
- you need Chinese-English bilingual dialogue capabilities
- you want to study or reproduce SFT and plugin-augmented LLM training
- you have a single GPU and need int4/int8 quantized inference

## When to avoid
- you need state-of-the-art reasoning quality comparable to modern frontier LLMs
- you want a managed API without hosting your own GPU infrastructure
- you need commercial use of the model weights without checking the AGPL-3.0 model license
- you need long-context or multimodal capabilities

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-inference, llm-training, chatbot, nlp
- domain: large-language-models, artificial-intelligence, chatbots
- platform: python
- tags: conversational-ai, tool-augmented-llm, quantization, sft-data, open-source-model, fudan-university, natural-language-processing, gpu, linux

## Member repositories
- OpenMOSS/MOSS (main) score 68

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:52.089773+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 2026-08-29T17:14:15.175547+00:00, confidence not recorded.
  - readme: https://github.com/OpenMOSS/MOSS (fetched 2026-08-28T04:10:52.089773+00:00, sha 09b7f7d70601)
  - homepage: https://txsun1997.github.io/blogs/moss.html (fetched 2026-08-29T08:11:44.355251+00:00, sha 0c3ec03df31d)
- Data as of 2026-08-30T08:39:29.467469+00:00.