TellWorldAI World Archive

Home›Companies

OrganizationCompany

Baichuan Intelligence

Baichuan Intelligence (百川智能) is a Beijing AI company founded in April 2023 by Wang Xiaochuan. It released open-source language models from 2023 and later focused on applying AI to healthcare.

Founded 2023Website baichuan-ai.com Era Generative AI for everyone

2023 · 4 20232026
Events per year

What it is

Baichuan Intelligence (百川智能) is a Beijing-based AI company. Its legal name is Beijing Baichuan Intelligent Technology Co., Ltd. The company builds large language models and related products, and it has been known for open-source releases as well as for a closed model line and an AI assistant. Its name comes from the phrase 'a hundred rivers'.[1]

Origins

The company was founded on 10 April 2023 by Wang Xiaochuan, the founder and former chief executive of Sogou.[1] Wang's career is described in his own entry; here it is enough to note that he founded the company and led it in its early years. In 2023 Baichuan had 170 employees, and it started with US$50 million in seed capital.[1]

The company's first open-source model, Baichuan-7B, came in June 2023, and Baichuan 2, in 7B and 13B sizes, followed on 6 September 2023.[1][2] In October 2023 Baichuan raised US$300 million from investors including Tencent, Alibaba Group, Shunwei Capital and Xiaomi.[1]

Main features

Baichuan's model line has moved in several directions. Its open-source releases were general language models. In January 2024 the company launched Baichuan 3, a closed model with more than one hundred billion parameters.[3]

In May 2024 the company unveiled Baichuan 4 and its first AI assistant, Baixiaoying (百小应). The company said it aimed to build an AI "super app", meaning a single application that would bring together many services.[4]

The company also has a medical line. Baichuan-M1 was released in January 2025. Baichuan-M2-32B is described by the company as the world's strongest open-source medical model.[6] Baichuan-M3 was released on 13 January 2026.[6][7] Its GitHub organisation also hosts the omni-modal model Baichuan-Omni-1.5, the speech model Baichuan-Audio and official MCP servers. MCP, the Model Context Protocol, is a standard way for AI models to connect to outside tools and data.[8]

Funding and direction

On 25 July 2024 the company announced a round of about 5 billion yuan (US$687.6 million). The investors included Alibaba, Tencent, Xiaomi and state-backed funds, and the round valued the company at more than 20 billion yuan.[5] Investors dubbed Baichuan one of China's six "AI Tiger" start-ups.[1]

In March 2025 the company announced a restructuring of its teams to focus on applying AI to healthcare.[1] This change is visible in its later releases, which are mostly medical models.

Where it stands

Baichuan has moved from general open-source language models toward a focus on medicine, while keeping a presence in speech, omni-modal models and developer tools. The company has released both open-source and closed models, so its approach mixes the two.

The company's claims, especially about benchmark performance and the strength of its models compared with others, come largely from its own announcements. Readers who want to judge them should look for independent evaluations. The company's history, from a founding in 2023 to a medical focus in 2025 and 2026, is the most solid part of the record.

CapabilitiesLarge language models

TechniquesFoundation models

ApplicationsBiomedicine

Ecosystem & regionOpen sourceChina

20261

  1. ReleaseBaichuan releases the open Baichuan-M3 medical model

    On 13 January 2026 Baichuan released Baichuan-M3, a new-generation open-source medical-enhanced LLM, at a media briefing.

20253

  1. ReleaseBaichuan open-sources the Baichuan-M2 medical model

    In August 2025 Baichuan open-sourced Baichuan-M2-32B, its second medical-enhanced model, which it called the world's strongest open-source medical model.

  2. MilestoneBaichuan restructures to focus on AI for healthcare

    In March 2025 Baichuan disbanded its financial-industry enterprise team and restructured to focus on AI for healthcare, later releasing the open medical models Baichuan-M2-32B and Baichuan-M3.

  3. ReleaseBaichuan launches Baichuan-M1-preview and open-sources Baichuan-M1-14B

    On 24 January 2025 Baichuan released Baichuan-M1-preview and open-sourced Baichuan-M1-14B, a smaller model it described as the first open-source LLM trained from scratch for medicine.

20243

  1. FundingBaichuan raises 5 billion yuan from Alibaba, Tencent and Xiaomi

    In July 2024 Baichuan announced a round of about 5 billion yuan from investors including Alibaba, Tencent and Xiaomi, valuing the company at more than 20 billion yuan.

  2. ReleaseBaichuan launches Baichuan 4 and the Baixiaoying assistant

    On 22 May 2024 Baichuan unveiled Baichuan 4 and its assistant Baixiaoying in Beijing, presenting the assistant as a step toward an AI super app.

  3. ReleaseBaichuan releases Baichuan 3, a hundred-billion-parameter model

    On 29 January 2024 Baichuan released Baichuan 3, a closed model with more than one hundred billion parameters, and reported its own benchmark results against GPT-4.

20234

  1. ReleaseBaichuan releases the open Baichuan 2 models

    On 6 September 2023 Baichuan released Baichuan 2, its next-generation open model family in 7B and 13B sizes, trained on 2.6 trillion tokens and published with intermediate checkpoints.

  2. PolicyChina publishes the Interim Measures for the Management of Generative AI Services

    On 13 July 2023 China's Cyberspace Administration published the Interim Measures for the Management of Generative AI Services, which took effect on 15 August 2023. They apply to generative AI services offered to the public in mainland China.

  3. ReleaseBaichuan open-sources Baichuan-13B

    In July 2023 Baichuan, a Chinese start-up founded by Sogou founder Wang Xiaochuan, released Baichuan-13B as open models. Commercial use was allowed after registration.

  4. ReleaseBaichuan releases its first open model, Baichuan-7B

    Baichuan, a Chinese AI startup, released Baichuan-7B as an open-source, bilingual Chinese-English language model about two months after its founding.

Models and products over time

9
  1. 2023
    • Jun 2023 Baichuan-7B Model

      Baichuan-7B is an open-source, 7-billion-parameter language model for Chinese and English released in June 2023, the first model from Baichuan, trained on about 1.2 trillion tokens and evaluated on Chinese and English benchmarks.

    • Jul 2023 Baichuan-13B Model

      An open-source 13-billion-parameter language model from Baichuan, released in Base and Chat versions in July 2023, trained on 1.4 trillion tokens and benchmarked against Meta's LLaMA-13B.

    • Sep 6, 2023 Baichuan 2 Model

      Baichuan 2 is the second-generation open model family from Baichuan, released on 6 September 2023 in 7B and 13B sizes. Commercial use required an application and was limited by user numbers.

  2. 2024
    • Jan 29, 2024 Baichuan 3 Model

      Baichuan 3 is a closed language model with more than one hundred billion parameters, released by Baichuan on 29 January 2024. The company reported that it matched or beat GPT-4 on several Chinese benchmarks and on Chinese medical tasks.

    • May 22, 2024 Baichuan 4 Model

      Baichuan 4 is a closed large language model released by the Chinese company Baichuan on 22 May 2024, the company's first multimodal model and the engine behind its first AI assistant, Baixiaoying.

    • May 22, 2024 Baixiaoying Product

      Baixiaoying is Baichuan's first AI assistant app, launched on 22 May 2024 and powered by its Baichuan 4 model. It is built around search, including multi-round search, and document reading, and Baichuan has framed it as a step towards an AI "super app".

  3. 2025
    • Jan 24, 2025 Baichuan-M1 Model

      Baichuan-M1 is Baichuan's medical model line, released on 24 January 2025 with Baichuan-M1-preview and the open Baichuan-M1-14B, which Baichuan described as trained from scratch for medicine.

    • Aug 2025 Baichuan-M2 Model

      Baichuan-M2 is an open-source medical-enhanced model with 32 billion parameters, built on Qwen2.5-32B and released under Apache 2.0 for research and reference.

  4. 2026
    • Jan 13, 2026 Baichuan-M3 Model

      Baichuan-M3 is a new-generation open-source medical-enhanced language model with 235 billion parameters, built on Qwen3-235B-A22B and released under Apache 2.0 in January 2026.

Notes and sources

  1. Baichuanhttps://en.wikipedia.org/wiki/Baichuan
  2. baichuan-inc/Baichuan-7Bhttps://github.com/baichuan-inc/Baichuan-7B
  3. 号称中文评测超越 GPT-4,百川智能发布超千亿参数大模型 Baichuan 3https://www.ifanr.com/1574506
  4. Chinese start-up Baichuan debuts AI assistant powered by upgraded LLM, eyes AI super apphttps://www.scmp.com/tech/tech-trends/article/3263683/chinese-start-baichuan-debuts-ai-assistant-powered-upgraded-llm-eyes-ai-super-app
  5. Chinese AI start-up Baichuan raises US$700 million from Alibaba, Tencent, Xiaomihttps://sg.finance.yahoo.com/news/chinese-ai-start-baichuan-raises-093000785.html
  6. baichuan-inc/Baichuan-M2-32Bhttps://github.com/baichuan-inc/Baichuan-M2-32B
  7. 百川智能发布新一代开源医疗增强大模型 Baichuan-M3https://www.ithome.com/0/912/847.htm
  8. Baichuan Intelligent Technology (baichuan-inc)https://github.com/baichuan-inc

Cite this page: Baichuan Intelligence. TellWorld, https://www.tellworld.net/companies/baichuan. Updated 2026-11-08