Dots vs Muse is now the defining contest in consumer AI agents. OpenAI unveiled Dots at DevDay on 29 September 2026: always-on agents powered by GPT-6 Astra, drawn as colourful blobs with eyes and pitched as a “chief of staff”. Meta’s Muse, launched three weeks earlier, got there first with the same cute styling and a very different price. In a piece published on 1 October, The Verge’s Hayden Field argued that Dots is a deliberate shot at Meta that “strikingly fails to compete” in one key area: price.
Muse is free for most uses, with paid tiers for heavy users. Dots is only included with OpenAI’s most expensive personal plans and its premium business tier. Most of ChatGPT’s 1.2 billion users cannot use it at all.
This article compares the two agents on price, purpose, safety and reach, corrects one pricing detail in the Dots vs Muse coverage, explains why OpenAI is unlikely to give Dots away soon, and sets out what the Dots vs Muse fight means for UK users and businesses. For the launch itself, see our DevDay keynote recap.
Table of contents
- Dots vs Muse: What Each Agent Is
- The Price Gap in Dots vs Muse
- The Token Economics Behind Dots vs Muse
- Why OpenAI Cannot Simply Give Dots Away
- How OpenAI Is Framing Dots vs Muse
- Where Muse Leads in Dots vs Muse
- Dots vs Muse on Real Tasks
- Dots vs Muse in the Enterprise Market
- Dots vs Muse for UK Users and Businesses
- Dots vs Muse FAQ
- References and Further Reading
Dots vs Muse: What Each Agent Is
In Dots vs Muse, both products promise an assistant that works on your behalf in the background, but they come from different starting points. Muse grew out of Meta’s consumer apps; Dots grew out of ChatGPT’s work features and Codex.
| Feature | OpenAI Dots | Meta Muse |
|---|---|---|
| Launched | 29 September 2026 (DevDay) | 8 September 2026 |
| Cheapest way in | ChatGPT Pro at $100 a month | Free, with a weekly allowance |
| Model | GPT-6 Astra | Meta’s Muse models |
| Where it runs | Cloud computer; ChatGPT, Slack, Teams | Dedicated virtual machine per user |
| Integrations | 4,000+ plugins | Facebook, Instagram, WhatsApp, connectors |
| Texting the agent | Billed as coming soon | Supported |
| Pitch | Long, ambitious knowledge work | Everyday tasks and errands |
Dots in brief
OpenAI describes Dots as agents that operate continuously, with access to a cloud computer and more than 4,000 apps through its plugin ecosystem. Pro and Business Premium subscribers get one Dot included, and Enterprise customers can switch on a beta. Sam Altman said Dots were “inspired by the cool agents that we all watched in movies growing up”, and OpenAI’s demos even had them build virtual worlds, a pointed nod at Meta’s metaverse.
Muse in brief
Muse gives every user a free dedicated virtual machine that holds the agent and the user’s data. It handles reservations, purchases, calendars and appointments, and asks for approval before actions such as sending email or buying something. We described that set-up in our report on Muse’s dedicated VM tab.
The Price Gap in Dots vs Muse
Price is the heart of the Verge’s Dots vs Muse argument, and the numbers bear it out. They also need one correction.
Correcting the $20 figure
The Verge wrote that Dots is “restricted for now to users on its $20/month Pro plan and up”. ChatGPT’s $20 plan is Plus, and Plus does not include Dots. The cheapest personal route is Pro at $100 a month, followed by the $200 and $500 Pro tiers. Subscription Insider’s plan-by-plan breakdown lists Free, Go, Plus and Business Standard as having no Dots access. We set out how those Pro rungs compare in our analysis of ChatGPT Pro 500.
Muse’s three tiers
Muse is not entirely free either. As reported by Subscription Insider and Crypto Briefing, it is free up to 100 million tokens a week, with a $20 Power plan (500 million tokens a week) and a $100 Maximum plan (3 billion tokens a week). The real Dots vs Muse difference is where the paywall sits: Muse charges heavy users, while Dots charges everyone.
Monthly price for agent access, US dollars (published plan prices)
What the price means for reach
The practical result is a Dots vs Muse reach gap. Anyone with a Meta account in the United States can start using Muse today. ChatGPT has 1.2 billion users, by OpenAI’s count, but Dots is limited to the small slice who pay for Pro, Business Premium or Enterprise. OpenAI has not said how much work each plan includes or what extra capacity will cost.
The Token Economics Behind Dots vs Muse
Before asking why OpenAI charges, it helps to see what Meta is giving away. Muse’s plans are priced in tokens, the chunks of text a model reads and writes, so the allowances can be compared directly.
Muse’s weekly allowances
As reported, Muse Free includes 100 million tokens a week, Power 500 million and Maximum 3 billion. Dividing each paid allowance by its monthly price gives 25 million tokens a week per dollar on Power (500 million / $20) and 30 million on Maximum (3 billion / $100), so the bigger plan is slightly better value per token.
Muse weekly token allowance by plan, millions of tokens (as reported)
Why a free allowance is expensive
For a sense of scale, OpenAI’s API list price for GPT-6 Astra is $10 per million input tokens, so 100 million input tokens would cost $1,000 at that rate. Muse runs on Meta’s own models and its tokens are not directly comparable, but the sum shows why a free weekly allowance of that size is a costly promise, and why the Dots vs Muse gap is really a gap in who can afford to subsidise.
How OpenAI meters Dots instead
OpenAI has not published a token allowance for Dots. Its Pro plans are sold as multiples of the Plus allowance, roughly 5x on Pro 100, 10x on Pro 200 and 25x on Pro 500, which works out at $20 per unit of Plus usage on every rung. An always-on Dot draws on that same pool, which is one reason OpenAI says users will later be able to adjust a Dot’s speed or monthly workload.
Why OpenAI Cannot Simply Give Dots Away
The Verge’s explanation of the Dots vs Muse price gap is blunt: OpenAI “likely can’t afford it”. Altman did not dispute the cost.
“It uses a lot of compute”
“Dots are starting out as a premium product — it uses a lot of compute,” Altman said at a press Q&A after the keynote. “But you should of course expect us to do a mass-market thing for billions of people someday.” An always-on agent that keeps working through long tasks consumes far more computing than a chat that ends when you close the window.
OpenAI’s finances
OpenAI is reportedly seeking to raise $30 billion at a $1.4 trillion valuation, according to Bloomberg, and is nearing $70 billion in annual recurring revenue after growing 70% since the start of its third quarter, Axios reported. It also expects to spend $280 billion by 2030, according to the Financial Times, and faces investor pressure to tidy its finances before a public listing.
Meta’s advertising machine
On the other side of Dots vs Muse, Meta can subsidise Muse from one of the largest advertising businesses in the world. Advertising made up about 98% of Meta’s revenue, and second-quarter 2026 revenue was $60.8 billion, according to Crypto Briefing’s summary of analyst notes. A free agent that keeps people inside Facebook, Instagram and WhatsApp supports that business even before it earns a subscription.
What paying users could be worth
Analysts have started pricing the Muse subscription upside. Crypto Briefing reported a Goldman Sachs estimate that 10 million paying Muse users could generate $2.4 billion a year, which is 10 million x $20 x 12 months, i.e. everyone on the Power plan. Wedbush projected $5 billion in subscription revenue in 2027.
How OpenAI Is Framing Dots vs Muse
OpenAI is not trying to win Dots vs Muse on price. It is trying to change the question from “which agent is cheaper” to “which agent can you trust with real work”.
A chief of staff, not an errand runner
Altman said Dots are built for delegating “ambitious pieces of work”, such as quality testing an unreleased product, website design, planning a team offsite, rebalancing a portfolio and creating an email campaign. OpenAI’s Holly Li said using a Dot was “like talking to a second brain”. The aim, Altman said, is to go beyond booking reservations and ordering meals, the classic uses for consumer agents like Muse.
Staying on task for longer
Asked how Dots differ from Muse and Instinct, Thibault Sottiaux, OpenAI’s head of core products, said a Dot is “capable of actually staying on track in very long periods of time”. “Specialist” Dots for accounting and legal analysis are due soon, and OpenAI is working with Microsoft to bring them into Agent 365.
Safety as a selling point
OpenAI also leaned on trust, treating the Dots vs Muse choice as a cybersecurity question as much as a product one. Codex product lead Alexander Embiricos said OpenAI wants the “most trustworthy, safe, and secure assistant”, powered by its “most aligned model”. The keynote introduced OpenAI Private Intelligence, a data framework with stronger business controls and zero data retention options. That pitch landed days after Muse gave a YouTuber’s address to a stranger on Facebook Marketplace, which we covered in our report on the Muse address leak.
“We might move a tiny bit slower”
Glen Coates, OpenAI’s head of app platform, put the contrast most directly. “We might move a tiny bit slower, but we also know that if people have Dots that go out there and YOLO-buy stuff on Facebook Marketplace, that’s just not something we want to put 1.2 billion users through,” he told The Verge. OpenAI is being deliberate, he said, about how agents use plug-ins, sign in, and pay for things.
Where Muse Leads in Dots vs Muse
For all OpenAI’s arguments, Muse holds Dots vs Muse advantages that are hard to answer with features alone.
Free has won before
The Verge makes the uncomfortable point that ChatGPT itself rose by being free and easy to use, even without signing in. Muse is now applying that same lesson to agents. A free product reaches people who would never compare plans, and habits formed in the first weeks tend to stick.
Meta’s own apps are the integration
Muse connects naturally with Facebook, Instagram and WhatsApp, apps that billions already use daily. Google’s Gemini Spark agent platform has a similar advantage with Google’s own services, the Verge notes. OpenAI can connect Dots to thousands of outside tools, but plug-ins are rarely as seamless as an in-house link.
Early downloads
Muse started fast. Sensor Tower data cited in our earlier coverage showed about 730,000 US downloads in Muse’s first few days, slightly ahead of the original Meta AI app’s 707,000 over five days, a gap of 23,000, or 3.3%. One later report put Muse above 2.5 million mobile downloads within 13 days, briefly ahead of ChatGPT in the free app charts.
Texting already works
Muse and Instinct already let you text your agent. OpenAI highlighted texting a Dot several times at DevDay, but the feature is still billed as coming soon, and the voice demos on stage did not work. The scorecard below sums up where each side of Dots vs Muse leads today.
| Factor | Edge | Why |
|---|---|---|
| Price | Muse | Free tier versus $100 minimum |
| Reach | Muse | Any Meta account in the US |
| Long, complex work | Dots | Built for multi-hour tasks |
| Business tools | Dots | Slack, Teams, Agent 365, Enterprise controls |
| Social integration | Muse | Facebook, Instagram, WhatsApp |
| Safety record so far | Dots, for now | No public incident yet; Muse had the address leak |
Dots vs Muse on Real Tasks
The two companies’ own descriptions suggest how the agents divide the work. The examples below use the tasks OpenAI and Meta have named, not independent tests, so treat them as positioning rather than results.
Booking a table for Friday
This is Muse’s home ground. Reservations, purchases, calendars and appointments are the classic consumer agent jobs, and Muse asks for approval before it books or buys. Altman explicitly said Dots should go beyond this kind of errand, so in Dots vs Muse terms, a dinner booking is not where OpenAI wants to compete.
Planning a team offsite
OpenAI named offsite planning as a Dots task. It mixes research, comparison, budgets and messages to colleagues, and it runs over days rather than minutes. That suits an agent that, in Sottiaux’s words, stays “on track in very long periods of time”, and that can sit inside Slack or Teams where the team already talks.
Testing an unreleased product
Quality testing a product before launch is the clearest example of work Muse is not pitched for. It needs access to a build, a test plan and a written report. OpenAI’s Codex heritage shows here, and businesses comparing Dots vs Muse for this kind of job are really comparing Dots with developer tools.
Rebalancing a portfolio
Altman also listed portfolio rebalancing. That is the riskiest example on either list, because it moves money. Whichever agent you choose, keep a human approval step on every trade, and check what the provider says about liability before you let an agent near an investment account.
Dots vs Muse in the Enterprise Market
Framing Dots as a professional tool does more than avoid a price war with Meta. It points Dots at a market where paying is normal, and where the Dots vs Muse comparison matters far less.
Anthropic is the real target
Anthropic has led enterprise AI for years, the Verge notes, and OpenAI has put more resources into business customers over the past year. Chief financial officer Sarah Friar told CNBC that OpenAI’s enterprise business has doubled since July. A paid agent that lives in Slack and Microsoft Teams fits that push far better than a consumer app.
Business Premium as the bridge
Business Premium costs $125 a seat per month, or $100 billed annually, and includes a Dot. For many firms, that is a cheaper test of agentic work than hiring, and the comparison with a free consumer agent is irrelevant. In the business market, the Dots vs Muse question becomes Dots against Anthropic’s agents and Microsoft’s own.
The risk in the middle
The danger for OpenAI is the gap in between, where Dots vs Muse is decided by habit. If consumers form their agent habits on Muse while businesses buy from Anthropic and Microsoft, Dots could be squeezed. Altman’s promise of a mass-market version “someday” is an admission that the consumer seat is not yet his.
Dots vs Muse for UK Users and Businesses
For British readers, the Dots vs Muse debate is partly theoretical for now, because neither agent is fully available here.
Availability
OpenAI skipped the EEA, the UK and Switzerland for Pro subscribers at launch, according to its help centre, while Business Premium is available across supported ChatGPT regions. Muse is rolling out in the United States on iOS, Android and the web. A UK consumer cannot easily get either; a UK business can get Dots through Business Premium.
What to test first
If you trial an agent, start with work that is long but low-risk: research summaries, report drafts, inbox triage. Keep payments, contracts and customer messages behind human approval until you have seen how the agent behaves. Our page on AI employees and autonomous agents explains how we scope those trials.
Questions to ask any agent vendor
Ask where the agent runs, what data it keeps, how it signs in to your accounts, whether it can spend money, and how you stop it. Under UK GDPR, an agent that reads customer data is processing personal data, so you will need a lawful basis, a data processing agreement and a record of what it does. These questions matter more than which blob has the friendlier eyes, and they decide Dots vs Muse for a business far more than the monthly price does.
Dots vs Muse FAQ
What is the Dots vs Muse difference in one line?
Dots is OpenAI’s always-on agent on GPT-6 Astra, aimed at long knowledge work and included with paid plans. Muse is Meta’s consumer agent with a free tier, its own virtual machine for each user and tight links to Facebook, Instagram and WhatsApp.
How much does Dots cost?
There is no separate fee. One Dot is included with ChatGPT Pro ($100, $200 or $500 a month) and Business Premium, and Enterprise customers can enable a beta. Plus at $20 does not include Dots.
Is Muse really free?
Muse is free up to a weekly allowance of 100 million tokens, as reported. Heavier users can pay $20 a month for Power or $100 for Maximum.
Which is safer in Dots vs Muse?
OpenAI pitches Dots as the more trustworthy option, with privacy controls and zero data retention for businesses. Muse has already had a public mistake, sharing a user’s address with a stranger. Both are new, so neither has a long record.
Can I use Dots or Muse in the UK?
Not easily as a consumer. Dots on Pro skipped the UK at launch, and Muse is rolling out in the US. UK businesses can access Dots through ChatGPT Business Premium.
Will OpenAI make Dots free?
Altman said OpenAI expects to offer “a mass-market thing for billions of people someday”, but gave no date or price.
References and Further Reading
OpenAI’s new agent is a shot at Meta, but can it compete with free? (The Verge)
OpenAI starts Dots at $100 as AI agent pricing takes shape (Subscription Insider)
OpenAI rolls out Dots agents as Meta offers Muse free for most uses (TokenPost)
Analysts estimate 10M paying Meta Muse users could mean $2.4B a year (Crypto Briefing)
OpenAI launches Dots agents powered by GPT-6 Astra (TestingCatalog)
With Dots, OpenAI wants you to stop being afraid of its AI agents (Gizmodo)
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
