Zhipu Qingyan
Zhipu Qingyan (also known as ChatGLM) is the consumer AI chatbot from Z.ai (formerly Zhipu AI), a Tsinghua University spinoff that became the world's first publicly traded AI foundation model company when it IPO'd on the Hong Kong Stock Exchange in January 2026. Powered by the GLM model series (latest: GLM-5 with 744 billion parameters), it offers AI-powered dialogue, content creation, code generation, and agent capabilities. Available as a standalone app and via API, with over 25 million users. China-focused but expanding globally, particularly with developers.
Score generated by AI agents based on publicly cited evidence and reviewed by the project maintainer. Not independently validated.
Score History
Timeline events are AI-curated from public reporting. Score trajectory is derived from documented events.
Zhipu AI was founded as a Tsinghua University spinoff focused on knowledge graphs and academic AI research. With no consumer product and minimal commercial activity, enshittification vectors were almost entirely absent. The only concerns were standard Chinese tech labor norms and the baseline regulatory environment for technology companies operating in China.
With GLM-130B open-sourced and the pivot from knowledge graphs to LLMs complete, Zhipu began building a developer ecosystem. China's Algorithm Recommendation regulations (2022) and emerging AI governance framework increased regulatory complexity. The company remained pre-revenue with no consumer product, but growing institutional investment from Legend Capital and Qiming Ventures introduced shareholder pressure.
The launch of ChatGLM and the Qingyan consumer app marked Zhipu's transformation from research lab to commercial AI platform. Passing CAC security assessment required implementing mandatory political censorship in all outputs, immediately elevating algorithmic opacity. A $350 million fundraise from Alibaba and Tencent deepened institutional ties. The MaaS platform grew rapidly to 25 million users, but the consumer app prioritized engagement content over core AI functionality.
The U.S. Entity List designation in January 2025 was a watershed moment, making Zhipu the first Chinese LLM company sanctioned and sharply elevating regulatory risk. Simultaneous rapid growth brought new problems: the Qingyan app was flagged for unauthorized data collection, and censorship became more visible as GLM models refused queries about the Sichuan protests. Aggressive price cuts in the 2024 AI price war benefited developers but deepened cash burn, while sovereign AI infrastructure sales to Middle Eastern and African governments raised geopolitical conduct concerns.
Z.ai's January 2026 HKEX IPO made it the world's first publicly traded AI foundation model company, raising $558 million. The transition to public company brought immediate monetization pressure: GLM-5 launched with 30-60% price increases, computing constraints forced 80% cuts to new coding plan sign-ups, and user complaints triggered a 23% single-day stock crash. China's draft anthropomorphic AI regulations added new compliance burden for Zhipu's Lingxin companion subsidiary.
Alternatives
Open-source Chinese AI chatbot with strong reasoning and coding capabilities. Free to use with no subscription required. Easy switch — just download the app or use the web interface. Better benchmark performance on reasoning tasks, though Zhipu's GLM-5 is competitive on coding.
The most capable general-purpose AI chatbot globally, but not directly accessible in mainland China without a VPN. For users outside China or with VPN access, it offers broader capabilities and larger knowledge base. Free tier available; Plus subscription at $20/month.
ByteDance's flagship AI chatbot and China's most popular by active users (100M+ DAU). Comparable general assistant capabilities with strong multimodal features. Easy switch as a standalone app. Free to use with premium options available.
Dimensional Breakdown
Summaries below were written by AI agents based on the cited evidence. They are editorial interpretations, not independent research findings.
Dimension History
Timeline (30 events)
Zhipu AI founded as Tsinghua University spinoff
Professors Tang Jie and Li Juanzi from Tsinghua University's Department of Computer Science founded Zhipu AI in Beijing's Tsinghua Science Park, initially focusing on knowledge graphs and academic AI research. The company grew directly out of Tsinghua's Knowledge Engineering Group (KEG).
Zhipu co-founder helps unveil WuDao 2.0 trillion-parameter model
Tang Jie, Zhipu co-founder and BAAI researcher, helped develop the WuDao 2.0 model with 1.75 trillion parameters, at the time the world's largest multimodal AI model. The research directly informed Zhipu's pivot from knowledge graphs to large language models, establishing the GLM architecture foundation.
China's algorithm recommendation regulations take effect
The CAC's Internet Information Service Algorithmic Recommendation Management Provisions, the world's most extensive algorithm regulation, took effect requiring transparency over recommendation algorithms, user opt-out controls, and mandatory algorithm filing. Providers with 'public opinion attributes' faced security assessments. Though Zhipu had no consumer product yet, the provisions established the regulatory framework that would govern its future AI services.
GLM-130B bilingual model open-sourced
Zhipu open-sourced GLM-130B, a 130-billion-parameter bilingual (Chinese/English) pre-trained model that outperformed GPT-3 175B on LAMBADA benchmarks. Released with full weights, training logs, and code, it demonstrated Zhipu's commitment to open-source AI development and attracted significant global attention.
Series B funding from Legend Capital and Qiming Ventures
Zhipu AI raised hundreds of millions of yuan in a Series B round from Legend Capital, Qiming Venture Partners, and Tsinghua Holdings. The funding supported completion of the first 100-billion-parameter GLM model and began the transition from a pure research lab to a commercially-oriented venture, introducing institutional investor governance expectations while maintaining academic founding principles.
China adopts Deep Synthesis Provisions for AI content
The CAC, MIIT, and Ministry of Public Security jointly adopted regulations on deep synthesis technology requiring AI service providers to establish algorithm review, ethics review, content monitoring, and personal information protection mechanisms. Service providers with 'public opinion attributes or social mobilization capacity' faced mandatory algorithm registration and security assessments. Taking effect January 10, 2023, these provisions further tightened the regulatory framework constraining Zhipu's planned consumer AI products.
ChatGLM-130B and ChatGLM-6B launched simultaneously
Zhipu launched ChatGLM-130B as a consumer-facing chatbot at chatglm.cn and simultaneously open-sourced ChatGLM-6B, a 6.2-billion-parameter model enabling local deployment on consumer GPUs via INT4 quantization. ChatGLM-6B was pre-trained on 1 trillion tokens equally from English and Chinese, attracting over 10 million downloads.
China enacts Interim Measures for Generative AI Services
The CAC, jointly with six other Chinese authorities, issued binding regulations requiring generative AI providers to complete security assessments, file algorithms, and ensure outputs align with 'core socialist values.' This was the world's first binding regulation on generative AI, mandating political content filtering for all Chinese AI chatbots including Zhipu's products.
Zhipu among first batch approved for public AI services
Zhipu AI was among 11 companies in the first batch to pass CAC security assessment and algorithm filing, gaining approval to offer generative AI services to the public. This enabled the consumer launch of the Qingyan chatbot app, but also formalized the company's obligation to implement mandatory political censorship in all outputs.
Zhipu Qingyan consumer AI assistant launched
Zhipu launched Qingyan, its consumer-facing AI assistant based on ChatGLM2, with capabilities including general Q&A, creative writing, and code generation. Available as a standalone app and WeChat plugin, it rapidly grew to over 25 million users by year-end, positioning Zhipu in direct competition with Baidu's Ernie Bot and ByteDance's Doubao.
Zhipu secures $350 million from Alibaba and Tencent
Zhipu AI raised approximately 2.5 billion yuan ($350 million) in a funding round led by Alibaba and Tencent, with participation from Meituan, Ant Group, Xiaomi, and HongShan. The round valued Zhipu at approximately $2.8 billion, making it one of China's best-funded AI startups alongside MiniMax and Moonshot AI.
Zhipu workforce scales past 800 as $350M round closes
Following the $350 million fundraise from Alibaba and Tencent, Zhipu AI expanded to over 800 employees, making it China's largest AI startup by headcount. With 60-70% of staff in R&D, the rapid scaling tested the company's academic-origin governance culture as it transitioned from a Tsinghua research spinoff to a commercially-oriented platform serving 25 million users.
GLM-4 All Tools launched with autonomous agent capabilities
Zhipu released GLM-4, its next-generation foundation model with a 128K token context window, multimodal understanding, and the 'All Tools' autonomous agent feature enabling interaction with browsers, code interpreters, and image generators. Performance was characterized as comparable to GPT-4 on instruction following and long-context tasks.
Zhipu joins China's AI price war with 80% API price cut
Following DeepSeek's aggressive pricing on May 6, Zhipu slashed its GLM-3-Turbo API price by over 80%, offering rates as low as 0.1 yuan per million tokens and increasing free token allocations to 25 million. The move was part of an industry-wide 'race to zero' in Chinese AI pricing, subsidizing growth at the expense of profitability.
GLM-4 open-sourced with multimodal variant
Zhipu open-sourced GLM-4-9B and GLM-4V-9B models, supporting multiple languages and long-text processing. The visual model GLM-4V-9B offered multimodal capabilities comparable to GPT-4V. The open-source release under MIT license further reduced developer lock-in and contributed to the ChatGLM series reaching over 30 million global downloads.
Saudi Aramco's Prosperity7 leads $400 million funding round
Zhipu raised $400 million in a Series C round that included Saudi Arabia's Prosperity7 Ventures (Aramco's venture arm), valuing the company at approximately $3 billion. The Saudi investment deepened Zhipu's connections to Middle Eastern sovereign AI infrastructure projects and signaled its growing geopolitical significance as a Chinese AI champion.
Zhipu added to US Entity List as first sanctioned Chinese LLM company
The U.S. Commerce Department added Zhipu AI and 10 subsidiaries to the Entity List for allegedly 'advancing China's military modernization through advanced AI.' Zhipu received a special 'Footnote 4' designation applying restrictions to foreign-produced items containing U.S. technology. Beijing Lingxin Intelligent Technology, Zhipu's AI companion subsidiary, was also included.
Qingyan flagged for unauthorized data collection by Chinese regulators
China's National Cyber Security Information Centre flagged Zhipu Qingyan among 35 apps for illegally collecting user information beyond what users authorized, during an April 16 to May 15 detection period. The violation represented a form of covert data harvesting where the app collected personal data exceeding legitimate business needs.
RSF documents systematic censorship in Chinese AI chatbots
Reporters Without Borders published a comprehensive investigation revealing how state propaganda and censorship are structurally embedded in Chinese AI chatbots. The report documented how companies must implement 'security filtering' with hundreds of keywords per category across eight categories of political content, making chatbot outputs fundamentally constrained by invisible political rules.
Zhipu rebrands to Z.ai and releases GLM-4.5 open-source model
Zhipu AI rebranded internationally as Z.ai while releasing GLM-4.5 and GLM-4.5-Air under MIT license. GLM-4.5, trained on 15 trillion tokens with 355 billion parameters and 32 billion active, featured a hybrid reasoning mode and 96K output tokens. The rebrand accompanied expansion with offices in the U.S., U.K., and France, and a tenfold increase in overseas paid users.
GLM-4.5 refuses all queries about Sichuan protests
During protests in Jiangyou, Sichuan over a police response to a bullying case, Zhipu's GLM-4.5 immediately refused to answer any related questions, as documented by China Media Project. This demonstrated AI chatbots' role as a new instrument of real-time information control during breaking news events, suppressing public discourse alongside traditional social media censorship.
Mandatory AI content labeling takes effect in China
China's Measures for Labeling AI-Generated Synthetic Content took effect, requiring all generative AI content to carry both explicit labels (visible 'AI-generated' notices) and implicit labels (digital watermarks). Zhipu implemented compliance measures including explicit labels on Qingyan responses and hidden digital identifiers on all text, images, and videos.
GLM-4.7 open-sourced as coding-focused agent model
Z.ai released GLM-4.7, an open-source model optimized for coding workflows with 'Preserved Thinking' capability maintaining reasoning chains across turns. Priced at $3/month (or free locally), it achieved the highest score on interactive tool-use benchmarks among open-source models and gained significant U.S. developer traction, competing directly with Western coding tools priced at $200/month.
China drafts world's first anthropomorphic AI regulations
The CAC released draft 'Interim Measures for Anthropomorphic AI Interaction Services' for public consultation, directly affecting Z.ai's Lingxin subsidiary. The rules prohibit emotional manipulation, require addiction detection algorithms, mandate guardian consent for minors, and impose security assessments for services exceeding 1 million registered users.
Z.ai IPO on Hong Kong Stock Exchange raises $558 million
Z.ai became the world's first publicly traded AI foundation model company, listing on HKEX at HK$116.2 per share. The IPO raised HK$4.3 billion ($558 million), with Hong Kong public offering oversubscribed 1,159 times. The stock opened at HK$120 and closed at HK$131.5, giving a market value exceeding HK$57 billion. The company allocated 70% of proceeds to R&D.
Coding plan sign-ups restricted to 20% amid computing constraints
Z.ai limited new GLM Coding Plan sign-ups to 20% of prior daily levels due to computing infrastructure constraints, cutting new sales by 80%. Users had reported service degradation including response delays and rate limits since the IPO. The company issued public appeals for domestic and international computing resource providers, the first such call from a Chinese AI company.
GLM-5 launched with 30-60% price increases
Z.ai released GLM-5, a 744-billion-parameter Mixture-of-Experts model with 40 billion active parameters, while simultaneously raising Coding Plan subscription prices by 30-60% domestically and 67-100% internationally. First-time purchase discounts were eliminated. The stock surged 30% on the day. This marked the first significant price increase in China's LLM market after a year of subsidized pricing.
PNAS study documents systematic political censorship in Chinese LLMs
A peer-reviewed study in PNAS Nexus by Stanford researchers tested ChatGLM and three other Chinese models against 145 political questions. ChatGLM had a 10% refusal rate (lowest among Chinese models, versus 36% for DeepSeek), but still showed substantially shorter and less accurate responses on politically sensitive topics compared to Western models. The study confirmed structural censorship baked into model alignment training.
Shares tumble 23% on computing woes and user complaints
Z.ai's stock plummeted nearly 23% in a single session as user complaints about service quality persisted and the company issued unprecedented public appeals for computing resource partners globally. Users in China and overseas reported ongoing issues with the GLM Coding Plan including response delays. The stock rebounded 20% the following day but exposed infrastructure fragility.
AutoClaw launched as domestic OpenClaw alternative
Z.ai launched AutoClaw, China's first one-click-install local version of OpenClaw, integrating the proprietary Pony-Alpha-2 model with over 50 pre-installed skills. The product responded to Chinese regulators banning OpenClaw from banks and state agencies over security vulnerabilities. Shares rose 13%, extending post-IPO gains to over 300%.