DeepSeek Harness, DeepSeek’s open-source agent app, had 1.54 million daily active users on 4 November and 5.48 million monthly active users, its team said in an update on the evening of 5 November, as reported by IT Home. The count covers only users of DeepSeek’s official API, the team said.
Those users consume about 92 trillion tokens a day, according to the update: about 60 million tokens per daily user per day, by our arithmetic. The team also said it will build an official plugin marketplace and is preparing a support programme for plugin authors, without giving a date.
Business Pill · The Model and the Harness
An evergreen Business Engineer explainer on the difference between a model and the harness around it. It teaches the general idea only and does not discuss DeepSeek or this update.
The key insight: As we read it, DeepSeek is reporting a fast-growing agent product with a plugin economy around it: 1.54 million daily users on 4 November, 5.48 million monthly, about 92 trillion tokens a day and about 35,000 third-party plugins, by its own count of official-API users. That is about 60 million tokens per daily user per day by our arithmetic. The next step it names is an official plugin marketplace, with no date or terms yet.
DeepSeek Harness v0.1 went live on 13 August and the desktop version was released on 29 September, according to the update. Since launch the number of third-party plugins has grown to about 35,000, and about 60% of active users have made or installed at least one, the team said (our translation of IT Home’s Chinese report).
A graphic published with the update shows average daily users rising from 271,000 in the week from 30 September to 429,000 from 7 October, 665,000 from 14 October, 1.13 million from 21 October and 1.40 million from 28 October, then 1.54 million on 4 November. The graphic says each point is a seven-day average from the date shown and that 4 November is a single-day value. It puts the rise at +545% compared with the desktop release on 29 September.
The graphic marks the release of DeepSeek’s V4.1-Flash-1017 model on 17 October. The week starting 21 October averaged 70% more daily users than the week starting 14 October, the largest weekly rise in the series, by our arithmetic; the update does not say what drove it.
Daily users were about 28% of monthly users on 4 November, by our arithmetic (1.54 million of 5.48 million). Pandaily and other outlets relayed the figures; one headline put the rise at 468%, which matches 1.54 million against 271,000, but that compares a single day with a seven-day average, so the team’s own +545% uses a different base. The Harness repository on GitHub had 273,941 stars when we read it on 6 November, against 269,000 in our 2 November piece.
Pandaily summarised the product’s design as everything is a plugin: in its description, models, tools, skills, sessions, sandboxes, storage and the user interface are all assembled from plugins. IT Home described Harness at its August public beta as positioned against Claude Cowork.
The Plugin Marketplace Plan
The team said it will “build an official plugin marketplace” (our translation), improve the stability and compatibility of the plugin API and strengthen plugin security. Its planned author programme would offer token incentives, technical exchanges and work recommendations, with details to follow community feedback. It also invited start-ups and companies to bring products and services into Harness as plugins. No launch date, revenue share or fees were given.
As we read the graphic, the steepest week came after the V4.1-Flash-1017 release on 17 October: the seven-day average rose 70% from the week of 14 October to the week of 21 October, by our arithmetic. The update does not link the two.
As we read the base, the 468% figure in some headlines and DeepSeek’s own +545% measure from different points, and the latest number is a single day against earlier seven-day averages.
As we read the plugin numbers, about 60% of active users have made or installed a third-party plugin, which is why DeepSeek describes a marketplace and an author programme as its next steps.
The DeepSeek Harness team, as carried by IT Home (our translation)
FOR DEEPSEEK The update reports 5.48 million monthly users on its official API through Harness and about 92 trillion tokens a day, by its own count.
FOR PLUGIN AUTHORS The team says it is preparing token incentives, technical exchanges and work recommendations, with details to follow community feedback.
FOR READING THE FIGURES They are DeepSeek’s own, count official-API users only, and mix a single day with seven-day averages.
The framework says: “Value migrates model → harness → skill marketplace.”
As we read it, DeepSeek’s plan for an official plugin marketplace on top of its Harness follows that sequence. The framework does not discuss DeepSeek Harness or this update.
What Is Not Established
The figures are DeepSeek’s own and count only official-API users; they have not been independently verified, and they do not include people running Harness with other models. The update gives no revenue, no paid-user count and no launch date, fee or revenue share for the plugin marketplace. It does not say what drove the October rise. The 1.54 million figure is a single day and is not directly comparable with the seven-day averages before it.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
DeepSeek says its Harness agent app had 1.54 million daily and 5.48 million monthly users on 4 November, burning about 92 trillion tokens a day, roughly 60 million per daily user by our arithmetic. Daily users rose from a 271,000 weekly average at the start of October. The team plans an official plugin marketplace and an author programme, with no date or terms yet.
A note on sourcing. The user, token, plugin and marketplace figures are from the DeepSeek Harness team’s update of 5 November 2026 and its graphic, as carried by IT Home in Chinese; the English renderings are our translation. Pandaily’s report and the GitHub star count (read 6 November) are cited as named. The tokens per user, daily-to-monthly ratio and weekly growth are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
