Zhipu (Z.ai): The First Listed LLM Company and the GLM Agent Bet

The 2026 numbers, with their labels

• 2513 HKEX ticker, listed January 2026 — Reported as the first publicly listed large-language-model company.
• 744B GLM-5.2 total parameters — Mixture-of-experts flagship (June 2026), roughly 40B active per token, 1M-token context. Company-reported architecture.
• $1.40/$4.40 GLM-5.2 per million tokens — Provider pricing at publication — after two reported price increases in 2026.
• 2019 founded, out of Tsinghua KEG — Spun out of Tsinghua University's Knowledge Engineering Group.

The GLM line: open weights, agentic focus, rising prices

Zhipu's 2026 model cadence ran GLM-5 (February, a 744-billion-parameter mixture-of-experts flagship), GLM-5.1 (April, released open-source alongside a reported ~10% API price increase) and GLM-5.2 (June), which keeps the 744B architecture with about 40 billion active parameters per token and a million-token context. The company positions GLM-5.2 for long-horizon agentic work — multi-step coding above all — and media coverage reports it matching or beating US frontier models on several such benchmarks at a fraction of the cost.

The pricing direction is the notable part: while the sector's story is falling prices, Zhipu raised API prices twice in 2026 — a listed company visibly trading some growth for margin. At $1.40 input and $4.40 output per million tokens it still undercuts US frontier rates by a wide multiple.

AutoGLM: the phone-use bet on agents

Zhipu's distinctive agent product is AutoGLM, open-sourced in December 2025 and described in industry coverage as the first agent with phone-use capability: it operates real Chinese consumer apps — the Douyin and Meituan class — through their interfaces to complete tasks like ordering or drafting. Where Grok Bot gives an agent its own cloud computer and Manus gave it a virtual machine, AutoGLM points the agent at the device people already live on.

Our read: phone-use is the most consumer-exposed form of the digital-labor race, because it automates apps rather than workflows. If it works at scale, the affected cohort is not knowledge workers first — it is the app ecosystem's interaction layer itself. That is a different shockwave shape than the office-agent products this site has covered, and worth its own tracking.

What a listed lab changes for the record

Being listed makes Zhipu the first frontier lab whose numbers face securities-grade disclosure. Revenue, costs and risks now appear in filings rather than blog posts, and claims can be audited against them. With Moonshot reportedly sprinting toward its own Hong Kong listing, Zhipu's January IPO looks less like an outlier and more like the opening of a public-market era for Chinese frontier labs.

For this site's method — separating verified fact from company claim — that era is a structural upgrade. It is also a discipline test for the labs: public markets price misses, and the gap between benchmark claims and billable usage becomes visible quarterly.

What this special does not claim

This file separates sourced reporting from open questions. As of publication:

• Architecture figures and benchmark comparisons are company- or media-reported; none are independently verified by this site.
• We have not tested GLM models or AutoGLM hands-on for this file; it is a sourced company file, not a review.
• AutoGLM's phone-use capability claims reflect launch coverage; real-world reliability at scale is unestablished here.
• Financial details beyond the listing facts belong to Zhipu's own disclosures, which this file does not restate.

Related reading on this site

• The Model Price War — Zhipu's two price increases run against the sector's falling-price story this site tracks weekly.
• Kimi company file — The other open-weight-first Chinese lab — reportedly heading for the same exchange.
• Claude Academy: Who Bears the Cost of AI Fluency? — Anthropic has launched a free school for learning to work with AI. Its stated framework reaches beyond prompts into delegation, judgment and disclosure. That is a meaningful public resource—and it raises a harder question: when the maker of the disruption also issues the credentials for adapting to it, where does institutional responsibility end and individual responsibility begin?
• Meituan All-in AI: The Execution Costs — Going all-in on AI is easy to announce and hard to govern. The execution bill arrives where strategic urgency meets source provenance, merchant consent and incentives: the less time a team leaves for verification and reversal, the more expensive its speed becomes. This special separates the verified public record from two weak, single-source signals and treats the pattern as a governance problem—not proof that AI investment itself has failed.
• AI Token Monetization: Token Is the New Dollar — At Stripe Sessions, President of Technology and Business Will Gaybrick changed a demo app from a $2 flat fee to $3 per million tokens, then streamed stablecoin payments as each token was consumed. That sequence is more than a billing demo. It shows software moving from seats and monthly access toward metered intelligence: every unit of model work can carry a price, a margin, a fraud risk and a settlement event. Our thesis is that the token is becoming the dollar of AI software—a unit of account for machine work, not legal tender and not a replacement for the US dollar.
• Stripe × OpenRouter: Where Token Monetization Begins — Put the reported $8 billion price aside. Stripe, the payments leader that became a checkout layer for the internet, chose not to buy a model lab but the switchboard between AI applications and hundreds of models. That is the story. Stripe is betting that the most important layer of the AI economy will not only produce intelligence; it will turn intelligence into exchangeable value. Today an LLM token is a billing unit. Tomorrow a model-agnostic AI token could be held, transferred and settled like the generation of digital assets opened by BTC and ETH — representing not digital scarcity or blockchain gas, but a claim on usable intelligence. Axios reports an agreement above $8 billion, while neither company had publicly confirmed it at this update.
• Grok Bot: xAI Gives Every Agent Its Own Computer — Launched in early beta on August 11, 2026, Grok Bot turns the agent from a chat window into a teammate: each Bot gets a persistent cloud computer, signs into the tools you already use and keeps working while you are away. We have been using it since day one. The product is days old and public information is still thin, so this special leads with tested impressions and keeps mechanism facts second — separating what is verified, what is company-stated and what is our read.
• Qwen: The Open-Weight Leader Starts Charging for the Crown — Alibaba's Qwen is the most-downloaded open-weight model family in the world — by company count, more than three billion downloads and over half the open-source market. In August 2026 it shipped its biggest flagship yet, priced it far under US frontier rates, and put its open weights under a revenue-share license for the first time. This file collects the sourced record and our read on what the pivot means.
• DeepSeek: The Price Anchor Starts Moving — Eighteen months after the R1 moment made frontier-for-pennies the industry's reference point, DeepSeek promoted V4-Pro to general availability — and in the same week announced API price increases of up to 1,100% plus the sector's first peak/off-peak token pricing. The lab that anchored the price war is repricing. This file collects the sourced record and our read.
• MiniMax: The Multimodal Tiger That Doubled on Debut — MiniMax reached the Hong Kong exchange one day after Zhipu and doubled on its first day. But the reason it closes this series is not the listing — it is the product surface. Where the other five files cover text and agents, MiniMax ships video, speech and music at commodity prices, and that points the AI shockwave at a different cohort: creators. This file collects the sourced record and our read.
• A Tribute to Manus — The independent special that anchors this site's digital-labor storyline.

Try Z.ai

Zhipu's international products run under the Z.ai brand, with GLM chat and developer API access.

Open z.ai

Official link — no affiliate relationship. If Zhipu opens an affiliate program, this site will disclose it.

Sources and evidence boundaries

Listing and model facts are anchored to reported coverage; architecture and benchmark figures are labeled company-reported. Pricing is taken from listings current at publication.

1. Pandaily — Zhipu launches Hong Kong IPO — IPO with HK$3B in cornerstone commitments; listing coverage.
2. SCMP — Zhipu open-sources flagship, raises prices — Open-weight strategy and the reported price increases.
3. DataNorth — Zhipu releases GLM-5.2 — GLM-5.2 architecture, agentic positioning and benchmark claims.
4. Baidu Baike — AutoGLM — AutoGLM's phone-use agent record, including the December 2025 open-source release.

Cite this

anti-ai.app, “Zhipu (Z.ai): The First Listed LLM Company and the GLM Agent Bet”, https://www.anti-ai.app/specials/zhipu/ (2026-08-15 · 2026-08-17).

Republication policy