MiniMax: The Multimodal Tiger That Doubled on Debut

The 2026 numbers, with their labels

• +109% first-day pop, HKEX debut — Listed January 9, 2026 — one day after Zhipu — closing at HK$345 vs a HK$165 offer; about $13.7B valuation, ~$620M raised. Reported figures.
• 2K · 15s Hailuo 3.0 (H3) omni-modal video — Launched July 31, 2026: text/image/video/audio in, video with native stereo sound out. Company-stated capability.
• $7.80/min 2K video generation price — $0.13 per second at 2K per official pricing; a 768p tier at $0.09/s reported in closed beta.
• <$20M community-license revenue cap — H3 open weights promised under a MiniMax Community License: commercial use only below $20M revenue, with attribution; Base-layer checkpoints at 768P. Per launch coverage.

The listing: the second AI tiger, one day behind the first

MiniMax's Hong Kong debut on January 9, 2026 came one day after Zhipu's — and outdid it as theater: the stock closed 109% above the offer price, valuing the company around $13.7 billion on roughly $620 million raised. Together the back-to-back listings opened the public-market era for Chinese frontier labs that this series has kept returning to.

The consequence is the same one our Zhipu file describes: securities-grade disclosure. Two of the six labs in this series now publish audited numbers, and the gap between launch-day claims and billable reality becomes visible quarterly.

The product surface: agents plus a full media stack

MiniMax runs the widest modality spread in this series. The M-series carries its text and agentic LLM line — currently reported at M2.5-to-M3 — alongside a standalone MiniMax Agent product. Then comes the media stack the other five labs do not have: the Hailuo video line through Hailuo 3.0, Speech 2.6 for voice, and Music 2.0.

H3, launched July 31, 2026, is the flagship of that stack: an omni-modal generator that takes text, images, video and audio together and returns up to fifteen seconds of 2K video with native stereo sound, priced at $0.13 per second — $7.80 per minute of generated footage. Whatever one thinks of any single demo, video at commodity prices is now a shipping product line, not a research preview.

Conditional openness: the license pattern reaches video

H3's weights were promised under a MiniMax Community License: free commercial use only for organizations under $20 million in revenue, attribution required, and only the Base-layer checkpoints — capped at 768P — actually opening. The full 2K model stays behind the API.

That is the same shape as Qwen's revenue-share pivot in our previous file: open weights as ecosystem seeding below a revenue line, paid access above it. Two labs, two modalities, one conclusion — as the artifacts get more expensive to train, pure open weights are giving way to conditioned openness. The pattern now spans text and video.

On the shockwave timeline: this one points at creators

The other five files in this series mostly pressure knowledge work: agents that execute, models that code. MiniMax's stack points somewhere else — video production, voiceover, and music are creator-economy trades, and $7.80 per minute of 2K footage with sound is a direct price signal to that cohort.

Our read: if the task×cohort framework this site runs on is right, MiniMax is the lab most likely to generate the next wave of timeline entries outside the office — in editing suites, dubbing studios and stock-media libraries. What lands there is an evidence question, and the media stack's real production adoption is the number to watch.

What this special does not claim

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

• Listing figures are market reporting; model capabilities and prices are company-stated and not benchmarked by this site.
• The Community License terms follow launch coverage; final license text and the open-weights delivery date remain to be confirmed.
• We have not tested MiniMax products hands-on for this file; it is a sourced company file, not a review.
• Creator-cohort displacement is our editorial framing of a risk, not an observed outcome; the timeline records outcomes when they land.

Related reading on this site

• Zhipu company file — The first listed LLM company — one day ahead; together they bookend the public-market turn.
• Qwen company file — The other conditional-openness pivot: revenue lines are entering open-weight licenses across modalities.
• 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.
• Kimi: Moonshot AI's Open-Weight Sprint to the Frontier — In four months Moonshot AI shipped an open-weight 1-trillion-parameter workhorse, followed it with the reported 2.8-trillion-parameter Kimi K3, paused new paid consumer subscriptions when demand outran capacity, and set off toward a Hong Kong listing. This file collects what is sourced, what is company-reported, and our read on where it fits the timeline this site tracks.
• 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.
• A Tribute to Manus — The independent special that anchors this site's digital-labor storyline.
• Model Price Watch — Rolling ledger of frontier model price moves, updated as list prices change.

Try MiniMax

MiniMax's international products run at minimax.io, spanning the agent, Hailuo video, speech and music lines.

Open minimax.io

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

Sources and evidence boundaries

Listing facts anchor to market coverage of the January debut; product and pricing facts to launch coverage and official pricing; license terms are early reporting.

1. CNBC — MiniMax doubles in Hong Kong debut — The January 9, 2026 listing, first-day close and valuation context alongside Zhipu.
2. Hugging Face community — MiniMax H3 (Hailuo 3.0) explained — H3's omni-modal design, 2K/15s output and the Community License structure.
3. Segmind — MiniMax H3 release, open weights and API pricing — Per-second pricing, the 768p beta tier and Base-layer checkpoint scope.
4. MiniMax — official site — First-party product line reference for the M-series, Agent, Hailuo, Speech and Music.

Cite this

anti-ai.app, “MiniMax: The Multimodal Tiger That Doubled on Debut”, https://www.anti-ai.app/specials/minimax/ (2026-08-15 · 2026-08-17).

Republication policy