# SkyworkAI/Skywork

Skywork series models are pre-trained on 3.2TB of high-quality multilingual (mainly Chinese and English) and code data. We have open-sourced the model, training data, evaluation data, evaluation methods, etc.

Repository: https://github.com/SkyworkAI/Skywork
Canonical: https://ross.abutalabs.com/products/skywork
Language: Python
License: NOASSERTION
License Family: other
Topics: llm
Last push: 2025-03-07T04:44:47+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 10, release rhythm 35, longevity 75
- inputs: {"age_days": 1058, "days_push": 544, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1498, forks 148 (observed 2026-08-28T04:04:54.026898+00:00)

## What it is
The Skywork series of open-source large language models (13B Base, Chat, Math, and multimodal MM variants) pre-trained on 3.2TB of multilingual Chinese/English and code data, released with quantized versions for consumer GPUs. The project also open-sources the Skypile-150B Chinese web dataset along with evaluation data and methods.

## Use cases
- run a chinese-english llm locally on a consumer gpu
- download a large pretraining dataset of chinese web text
- use a 13b model fine-tuned for math word problems
- chat with an open-source llm good at creative writing
- do image question answering with an open multimodal model
- evaluate open llms on gsm8k and cmath benchmarks

## When to choose
- you need strong bilingual Chinese/English base or chat models with commercial-use terms
- you want reproducible training data and evaluation methodology alongside model weights
- you need quantized 13B models that fit consumer-grade GPUs

## When to avoid
- you need the smallest possible model or a permissive license like Apache/MIT
- you only want an inference server rather than model weights and datasets
- you require cutting-edge frontier-scale performance beyond 13B-class models

## Facets
- artifact type: dataset
- maturity: active
- function: machine-learning, llm-inference, llm-training, nlp
- domain: large-language-models, machine-learning, data-science
- platform: python, cross-platform
- tags: llm-weights, model-release, chinese-nlp, multimodal, open-dataset, skypile, quantization, consumer-gpu, natural-language-processing, gpu, linux

## Member repositories
- SkyworkAI/Skywork (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:54.026898+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 2026-08-30T04:33:04.187575+00:00, confidence not recorded.
  - readme: https://github.com/SkyworkAI/Skywork (fetched 2026-08-28T04:04:54.026898+00:00, sha 11ff5cfe31d9)
- Data as of 2026-08-30T08:39:29.467469+00:00.