# Langboat/Mengzi3

Repository: https://github.com/Langboat/Mengzi3
Canonical: https://ross.abutalabs.com/products/mengzi3
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2024-10-09T03:45:09+00:00

## Health v2 (maintenance only)
Score: 25/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 63
- inputs: {"age_days": 887, "days_push": 693, "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 1369, forks 16 (observed 2026-08-28T04:04:31.717477+00:00)

## What it is
Mengzi3 is a series of open-source Chinese-focused large language models (8B and 13B parameters) released by Langboat in Base and Chat variants. The repository serves as the hub for model weights, links to Hugging Face/ModelScope downloads, and usage documentation.

## Use cases
- download a chinese open-source llm
- run a chinese chat model locally
- fine-tune a chinese base model
- compare mengzi3 8b vs 13b
- self-host a chinese language assistant
- use a chinese llm for rag

## When to choose
- you need an open-weights LLM with strong Chinese language capability
- you want base and chat variants for fine-tuning or direct use
- you prefer Apache-2.0 licensed models

## When to avoid
- you need a primarily English-optimized model
- you lack GPU resources for 8B/13B inference
- you want a hosted API only without self-hosting

## Facets
- artifact type: learning-resource
- maturity: active
- function: llm-inference, llm-training, chatbot
- domain: large-language-models, artificial-intelligence
- platform: python, cross-platform
- tags: chinese-llm, open-weights, model-weights, langboat, huggingface, natural-language-processing, gpu

## Member repositories
- Langboat/Mengzi3 (main) score 25

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:31.717477+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 2026-08-30T04:40:57.181542+00:00, confidence not recorded.
  - readme: https://github.com/Langboat/Mengzi3 (fetched 2026-08-28T04:04:31.717477+00:00, sha 00e072d97808)
- Data as of 2026-08-30T08:39:29.467469+00:00.