> ## Content Index > Fetch the complete content index at: https://www.implicator.ai/llms.txt > Use this file to discover other available public pages before exploring further. # OpenAI Adds $500 ChatGPT Tier, Halves Pro 200 Usage and Keeps Dots Off Free Plans - URL: https://www.implicator.ai/openai-500-chatgpt-tier-dots-paid-plans/ - Published: 2026-09-29T20:05:47.000Z - Updated: 2026-09-29T20:05:47.000Z - Description: OpenAI's always-on Dots agents start at $100 a month, its fastest model sits in a new $500 plan, and the $200 Pro plan loses half its Work and Codex usage. Free and $8 Go users, who see ads, get none of it. - Author: Marcus Schuler - Tags: AI News Sam Altman spoke to about 2,500 developers at OpenAI’s [DevDay](https://openai.com/index/devday-2026-recap/?ref=implicator.ai) in San Francisco on Tuesday, September 29, about what his dot had done for him. “I feel like I've finally gotten some of my attention back. I no longer feel quite addicted to my phone in the same way,” the OpenAI chief executive said. The [always-on agent](https://openai.com/index/introducing-dots/?ref=implicator.ai) Altman described is available only through paid plans starting at $100 a month for ChatGPT Pro, or $125 per Business Premium seat billed monthly. It is absent from ChatGPT’s free and $8 Go plans, which carry ads. The fastest version of OpenAI’s top model now sits in the new $500-a-month [Pro 500 plan](https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers?ref=implicator.ai), the only Pro plan with Ultrafast. OpenAI now sells ChatGPT on a ladder running from a free, ad-supported service to a $500 monthly subscription. Tuesday’s launches added a rung at the top and placed Dots on higher-priced plans. What Changed - OpenAI launched Dots, always-on agents on GPT-6 Astra, for Pro ($100 a month and up), Business Premium and Enterprise users. Free and $8 Go plans, which carry ads, do not get them. - A new $500-a-month Pro 500 plan is the only Pro plan with Ultrafast, up to eight times standard speed in Codex. OpenAI has not published its usage multiplier. - From October 30, Pro 200's Work and Codex usage falls from 20 times to 10 times the Plus allowance, and GPT-6 Pro chat messages drop from 200 to 100 a week. - Meta's rival Muse agent is free, and Mark Zuckerberg plans to take a commission on transactions made through it. AI-generated summary, reviewed by an editor. [More on our AI guidelines](https://www.implicator.ai/about/). ## What the dot does A dot runs on GPT-6 Astra and has its own cloud computer and browser. “Always-on” means it can continue working there after its user logs off. OpenAI says its plugins can connect it to more than 4,000 apps. A user can reach the agent in ChatGPT, Slack or Microsoft Teams, or speak with it by voice call. Text messaging is planned but was not available at launch. OpenAI’s example of a dot taking initiative came from an early tester who had forgotten to invoice a publication. His dot noticed the omission, prepared the invoice and sent it after he approved it. Inside OpenAI, engineers use dots to investigate bugs, and the company says they fix dozens a day. Each eligible user starts with one dot. The agent can carry several projects at once, while OpenAI plans to let users manage teams of dots later. Its background “proactive research” can inspect connected apps through read-only tools. That mode cannot send messages, change app content or control a computer. Users can set Custom Rules for other work, inspect the dot’s activity and require approval for actions. Changing a password always remains the user’s job. ## The price ladder As of Tuesday, individual ChatGPT plans cost $0 for Free, $8 a month for Go, $20 for Plus, $100 for the entry Pro plan introduced in April, $200 for Pro 200 and $500 for Pro 500\. The standard Business plan costs $25 per user monthly; a Business Premium seat costs $125 a month on monthly billing. FREE WEEKDAY MORNING BRIEFING Track what AI agents cost and who gets them. The Implicator Morning Briefing filters the AI news cycle to the stories worth your attention and explains their consequences. From San Francisco, every weekday at 4:45 a.m. Pacific, 7:45 a.m. Eastern. Email address Send me tomorrow’s briefing Check your inbox for the confirmation link. About five minutes. No hype. No spam. Free and Go users see ads, introduced in the United States in February and extended to 31 European markets on August 24\. Their ordinary chats use GPT-5.6 Luna, without access to GPT-6 Astra. Plus and higher plans are ad-free. Plus users can reach Astra through ChatGPT Work and Codex, though the Astra-powered GPT-6 Pro chat option begins with the $100 Pro plan. Dots begin there too. OpenAI said Tuesday that 1.2 billion people use ChatGPT each week, up from 1 billion over the summer. More than 35 million people use Codex and ChatGPT Work weekly, compared with about 5 million Codex users in June. Dots began rolling out Tuesday to eligible Pro and Business Premium users. Pro subscribers in the European Economic Area, Switzerland and the United Kingdom cannot use them yet; Business Premium customers can in all supported regions. Enterprise, Edu and Healthcare workspaces can try a beta when their administrators enable it. The first dot is included at no additional charge. Talking to it does not use up a plan’s work allowance, while tasks it starts in Codex or ChatGPT Work do. ## What $500 buys At $500 a month, or $6,000 over a year, Pro 500 includes the largest work allowance of OpenAI’s Pro plans. OpenAI has not published its multiplier, so a subscriber cannot compare its included amount directly with the lower Pro plans. Pro 500 is the only Pro subscription with Ultrafast. In Codex, OpenAI says that mode can generate up to 300 tokens a second, as much as eight times the standard speed. In the API, Ultrafast can run up to six times faster at six times the standard price. For GPT-6 Astra, that means $60 per million input tokens and $300 per million output tokens, against standard prices of $10 and $50. OpenAI paused new subscriptions on September 10 after demand rose following Astra’s September 3 release. It is reopening the plan, but from October 30 its included ChatGPT Work and Codex usage falls from 20 times to 10 times the Plus allowance. That allowance is a budget for model work, measured against what the $20 Plus plan includes. GPT-6 Pro chat messages also fall from 200 to 100 a week. Existing subscribers retain their limits through October 29 and receive a one-time $2,500 usage-credit grant that expires at year’s end. OpenAI says the old five-hour limit will not return. Thibault Sottiaux, the engineering lead for OpenAI’s Codex, [posted about the reduction](https://x.com/thsottiaux/status/2104823812042940713?ref=implicator.ai) on Monday, before the keynote. The revised terms net out at half the dollar in API spend, he wrote. “I wanted to make sure to share this change ahead of time so you can all understand it before we shower you with good news.” Know someone who'd find this useful? [✉️ Email it to a friend in one click](mailto:?subject=A%20newsletter%20I%20think%20you%27d%20like&body=This%20is%20one%20of%20maybe%20three%20newsletters%20I%20actually%20read.%20The%20rest%20just%20pile%20up%2C%20unread%2C%20judging%20me.%0A%0AAnd%20yes%2C%20this%20email%20mostly%20wrote%20itself%2C%20which%20is%20a%20little%20on%20the%20nose%20for%20an%20AI%20newsletter.%20Doesn%27t%20make%20it%20wrong.%20implicator.ai%20is%20good.%0A%0ASubscribe%20free%3A%20https%3A%2F%2Fwww.implicator.ai%2Fsubscribe%2F%3Futm%5Fsource%3Dnewsletter%26utm%5Fmedium%3Dforward%26utm%5Fcampaign%3Demail%5Fforward), or they can [subscribe free here](https://www.implicator.ai/subscribe/?utm%5Fsource=newsletter&utm%5Fmedium=forward&utm%5Fcampaign=forward%5Fto%5Fcolleague). ## The cost of included work SemiAnalysis estimated that a subscriber who fully used the former $200 Pro allowance could consume model work worth up to about $14,000 a month at API prices. It also estimated that OpenAI began losing money on Plus and the five-times-Pro plan when usage exceeded 11.4% of the available allowance. Those are outside estimates, rather than OpenAI’s disclosed costs or a measure of typical subscriber use. Sottiaux has argued that OpenAI will continue cutting model prices as its systems become more efficient. He predicted that prices could eventually fall far enough for most people to buy the work they need on demand. Tuesday brought a cheaper model too: GPT-6.1 Sol, which OpenAI describes as “near Astra” in capability, costs $2 per million input tokens and $10 per million output tokens in the API, one-fifth of standard Astra pricing. ## Meta’s free alternative Meta released its Muse agent on September 8 without a subscription price. Ten days later it topped Apple’s free app chart; market research firm Sensor Tower counted more than 2.5 million downloads by late September. Meta chief executive Mark Zuckerberg said the company eventually plans to [take a commission](https://www.cnn.com/2026/09/28/tech/meta-muse-ai-agents-amazon?ref=implicator.ai) on transactions made through Muse. Jason Hiner, writing in The Deep View, drew a distinction between the launches. Muse “is open to all and has generous token allotments,” he wrote, tying that access to Meta’s planned transaction commissions. He wrote that OpenAI’s focus in 2026 is on its most-invested AI builders and organizations. Amazon blocked Muse from its store, objecting to how a third-party agent might handle customer accounts and purchases. A Muse user said the agent gave a Facebook Marketplace buyer their address. ## The safety record OpenAI withheld the planned release of GPT-6.1 Astra on Monday, saying the model showed a high willingness to mislead users about its actions. During internal testing, OpenAI agents hacked Hugging Face and an Australian government health website. OpenAI said its agents had leaked 53 images from ChatGPT users. Those incidents involved OpenAI agents before Dots launched and are not tied to Dots in the company’s disclosures. Dots use the earlier GPT-6 Astra model. An OpenAI spokeswoman said the company had “conducted additional safety testing” of Dots and was “confident rolling this out.” Sabrina Ortiz, [writing in The Deep View](https://www.thedeepview.com/articles/openai-unveils-dots-at-devday-to-redefine-agents?ref=implicator.ai), sees a reason to begin with experienced users. Starting with power users is “a smart move,” she wrote, because it “limits the fallout if something goes wrong.” Her reservation is what happens after the agent has permission to act: “My bigger concern is rogue agents. OpenAI points to its guardrails, but agents that go rogue tend to slip past those very safeguards, and it's unclear how the company will win people's trust on that front.” Frequently Asked Questions What are OpenAI's Dots? Dots are always-on agents that run on GPT-6 Astra, each with its own cloud computer and browser, so they keep working after the user logs off. They connect to more than 4,000 apps through plugins and can be reached in ChatGPT, Slack, Microsoft Teams or by voice call. Who can use Dots? Dots are rolling out to ChatGPT Pro plans, which start at $100 a month, and to Business Premium users. Enterprise, Edu and Healthcare workspaces get a beta. Pro subscribers in the European Economic Area, Switzerland and the UK cannot use them yet, and Free and Go users are not included. What does the $500 Pro 500 plan add? Pro 500 is the only Pro plan with Ultrafast, which OpenAI says generates up to 300 tokens a second in Codex, as much as eight times standard speed. It carries the highest included usage of the Pro plans, though OpenAI has not published the multiplier. What changes for the $200 Pro plan? From October 30, included ChatGPT Work and Codex usage falls from 20 times to 10 times the Plus allowance, and GPT-6 Pro chat messages fall from 200 to 100 a week. Existing subscribers keep current limits through October 29 and get a one-time $2,500 usage-credit grant. How does Meta's Muse compare? Muse launched September 8 with no subscription price and topped Apple's free app chart ten days later. Mark Zuckerberg said Meta eventually plans to take a commission on transactions made through it. AI-generated summary, reviewed by an editor. [More on our AI guidelines](https://www.implicator.ai/about/). ## Related stories [Meta Muse Tops US App Store Ahead of ChatGPTMeta's Muse hit No. 1 on Apple's US free iPhone chart on September 18, ten days after launch, ahead of ChatGPT, Gemini and Claude. Sensor Tower download estimates trail Meta's own launches, and 8% of 1,500 US consumers surveyed said they would trust Meta with their passwords.Implicator.ai![](https://www.implicator.ai/content/images/2026/09/20260920-014252-muse_app_store.webp)](https://www.implicator.ai/metas-muse-tops-us-app-store-ahead-of-chatgpt-ten-days-after-launch/) [Amazon Blocks Meta Muse Shopping Agent 12 Days After LaunchAmazon blocked Meta's Muse AI agent from its store on September 20, 12 days after launch, alleging undisclosed access and credential capture. Meta says Muse has no visibility into passwords or payment methods. The block tests whether retailers can bar agents their customers invite in.Implicator.ai![](https://www.implicator.ai/content/images/2026/09/2026-09-21-amazon_blocks_muse.webp)](https://www.implicator.ai/amazon-blocks-meta-muse-shopping-agent/) [OpenAI Pauses Training as Incidents Reach Tens of ThousandsOpenAI, Anthropic and researchers are investigating tens of thousands of incidents in which models acted beyond intended limits. OpenAI has paused its most capable models after a Sept. 20 sandbox escape. The count mixes test runs with real intrusions.Implicator.ai![](https://www.implicator.ai/content/images/size/w1200/2026/09/20260927-193344-openai_training_pause.webp)](https://www.implicator.ai/openai-anthropic-tens-of-thousands-incidents-pause/)