Langboat 澜舟科技 (Mengzi)
Mengzi language models plus a new enterprise digital-worker OS
Overview
Langboat builds the Mengzi family of Chinese-strong language models (1B to 100B+ parameters) and, since June 2026, LangClaw — an enterprise agent OS for building and managing task-executing digital workers. Backed by Sinovation Ventures; clients span finance and telecom.
What you can produce with Langboat 澜舟科技 (Mengzi)
- Mengzi family of Chinese-strong LLMs (1B to 100B+ parameters)
- Open-source Mengzi3 (8B/13B, Apache 2.0, Llama architecture)
- LangClaw enterprise agent OS for orchestrating digital workers (launched June 2026)
- Langzhou Zhihui: AI meeting transcription, summarization and Q&A
- Retrieval-augmented enterprise knowledge-base product
- Backed by Sinovation Ventures (Kai-Fu Lee)
- Clients include CICC, Huawei, China Mobile, China Unicom, Huaxia Fund
ASEAN Perspective
Langboat 澜舟科技 (Mengzi) in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Langboat's edge is real: its Mengzi models topped CLUE benchmarks back in 2021, before most of today's LLM startups existed, and the open Mengzi3 (Apache 2.0, up to 13B parameters) still holds a respectable 1.4k GitHub stars. The company's June 2026 LangClaw launch — an enterprise 'digital worker' OS for orchestrating agents — is a sensible next step for an NLP shop that's been quietly building knowledge-base and meeting-transcription products for years rather than chasing consumer chatbot hype. A brokerage deployment case is concrete: 100,000+ research reports parsed with a claimed 95.51% answer accuracy. The catch is scale — Langboat just closed a Pre-A+ round in a market where rivals raise nine-figure rounds, and LangClaw is brand new with no independent track record yet. Good pick for Chinese-language enterprise NLP with real benchmark pedigree; wait for more field reports before betting on LangClaw specifically.
What people say
Langboat Technology (澜舟科技) was founded in Beijing in June 2021, incubated by Kai-Fu Lee's Sinovation Ventures, making it one of the earlier Chinese startups to bet specifically on large language models rather than narrower NLP tools. Its Mengzi (孟子) model family started with a 1-billion-parameter model that beat larger competitors to top the CLUE Chinese-language-understanding leaderboard in 2021 — a credential the company still leans on. Mengzi has since scaled up: Mengzi3, an 8B/13B model built on the Llama architecture and released under Apache 2.0 (free for commercial use), has picked up 1.4k stars on GitHub, unusually high engagement for a Chinese enterprise-AI vendor's open release. A Mengzi GPT chat model followed in 2023, with the company giving developers 4 million free tokens to try it.
The product line spans four tiers by the company's own framing — general-purpose Mengzi models, industry-specific models, task models and agents — surfacing in concrete products like Langzhou Zhihui (澜舟智会), an end-to-end meeting-analysis tool doing transcription, summarization and Q&A over recordings, and a knowledge-base product combining retrieval-augmented generation with the Mengzi models for enterprise document search.
The newest and biggest move is LangClaw, an enterprise agent operating system the company unveiled in April 2026 and opened publicly on June 10, 2026 to mark its fifth anniversary — pitched as shifting enterprise software from "people operating systems" to "people directing AI to execute." Financing has kept pace: Langboat closed a Pre-A+ round led by Beijing Zhongguancun Science City Company, with Strado Capital and Sinovation Ventures following on, bringing total raised to several hundred million RMB within roughly a year.
Client-side evidence is more concrete than most vendors in this space provide. One brokerage-firm deployment of Langboat's investment-research tools reportedly parsed more than 100,000 research reports and extracted meeting notes from 3,000+ audio/video files, with the company citing a 95.51% answer accuracy and 99.21% relevance score for its Q&A layer — numbers from the vendor's own case study rather than an independent audit. Named clients include CICC, Huawei, China Mobile, China Unicom, Tonghuashun and Huaxia Fund. No G2/Capterra presence was found; user feedback outside vendor case studies and Chinese trade press is limited.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Langboat 澜舟科技 (Mengzi)'s official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Langboat 澜舟科技 (Mengzi) unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Langboat 澜舟科技 (Mengzi) directly →
Spotted something out of date? Suggest an update →
Langboat 澜舟科技 (Mengzi) in the news
More in Business & Finance
