# baichuan-inc/Baichuan-7B

A large-scale 7B pretraining language model developed by BaiChuan-Inc.

Repository: https://github.com/baichuan-inc/Baichuan-7B
Canonical: https://ross.abutalabs.com/products/baichuan-7b
Homepage: https://huggingface.co/baichuan-inc/baichuan-7B
Language: Python
License: Apache-2.0
License Family: permissive
Topics: artificial-intelligence, ceval, large-language-models, natural-language-processing, mmlu, chatgpt, gpt-4, huggingface, llama, chinese
Last push: 2024-07-18T14:23:01+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 84
- inputs: {"age_days": 1176, "days_push": 776, "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 5649, forks 501 (observed 2026-08-28T04:09:26.998690+00:00)

## What it is
Baichuan-7B is an open-source, commercially usable 7-billion-parameter pretrained language model built on the Transformer architecture, trained on roughly 1.2 trillion tokens of bilingual Chinese-English data with a 4096-token context window. The repository provides the model code, inference and finetuning support via Hugging Face Transformers, and evaluation scripts for benchmarks like C-Eval and MMLU.

## Use cases
- run inference with an open-source chinese-english 7B language model
- finetune a commercially licensed alternative to LLaMA
- evaluate a 7B model on C-Eval and MMLU benchmarks
- build chinese-language text generation applications
- use a pretrained LLM as a base for downstream nlp tasks

## When to choose
- you need a permissively licensed (Apache-2.0) 7B model for commercial use
- your application is Chinese-focused and needs strong C-Eval performance at 7B scale
- you want a Hugging Face Transformers-compatible base model for finetuning

## When to avoid
- you need the newest or best-performing Baichuan model - Baichuan 2 supersedes this one
- you need instruction-tuned chat behavior out of the box, since this is a base pretrained model
- you lack GPU resources to run a 7-billion-parameter model

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-inference, llm-training, machine-learning, nlp
- domain: large-language-models, artificial-intelligence
- platform: python
- tags: pretrained-language-model, chinese-english-bilingual, transformer, 7b-parameters, huggingface-transformers, commercial-use-allowed, superseded-by-baichuan2, natural-language-processing, gpu, linux

## Member repositories
- baichuan-inc/Baichuan-7B (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:26.998690+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 2026-08-29T17:55:16.887801+00:00, confidence not recorded.
  - readme: https://github.com/baichuan-inc/Baichuan-7B (fetched 2026-08-28T04:09:26.998690+00:00, sha 805993864098)
  - homepage: https://huggingface.co/baichuan-inc/baichuan-7B (fetched 2026-08-29T08:49:47.723320+00:00, sha ba9bf92ad46e)
  - site_page: https://huggingface.co/docs (fetched 2026-08-29T08:49:47.732801+00:00, sha bdec26667b98)
  - site_page: https://huggingface.co/docs/inference-providers (fetched 2026-08-29T08:49:47.737163+00:00, sha 8a5d0f819473)
  - site_page: https://huggingface.co/docs/hub/model-cards (fetched 2026-08-29T08:49:47.739994+00:00, sha 60ded09a56b0)
  - site_page: https://huggingface.co/pricing (fetched 2026-08-29T08:49:47.734749+00:00, sha de6b7a178be5)
  - site_page: https://huggingface.co/huggingface (fetched 2026-08-29T08:49:47.742060+00:00, sha 1af65c1a389d)
- Data as of 2026-08-30T08:39:29.467469+00:00.