# ChatGPT Dots (OpenAI)

Feedback Bench, personal agents, built 2026-10-19, window 2026-08-31 to 2026-10-18. Web page: https://feedbackbench.com/#/agent/dots

| Measure | Value |
|---|---|
| Rank | 8 of 9 (rank range 8–8) |
| Feedback Score | 39.2 (95% interval 38.2–40.5) |
| Popularity | 0.746 (share of voice 9.52%) |
| Customer love | 0.379 (95% interval 0.366–0.393) |
| Place | Popularity 5th of 9 × Customer love 9th of 9 |
| Top quadrant | no |
| Users | 10842 |
| Posts counted | 14581 |
| Posts that judge the agent | 4304 |
| Criteria ranking high / low | 1 / 31 of 53 |

## The brief

Written by large language models from this page's numbers and the posts it quotes, then checked: every number in the text is in the build. Interpretation, not measurement.

People like talking to ChatGPT Dots, but it lets them down when it has to do the work. Users choose it for voice conversations, which make up 7% of what they rate, 5.5 times the rate for other agents. They also choose it for a usage allowance that lasts. The work itself pulls it down. It ranks last of 9 on getting tasks done, and posts describe slow replies and jobs handed off to Codex. It is 8th of 9 overall and last on customer love. When users compare it with Grok Bot and Muse, those agents mostly come out ahead. It leads OpenClaw.

### Why people choose it

- **Voice calls feel natural and useful** ([Good voice conversations](https://feedbackbench.com/criteria/ask.voice.md)). 5.5×: 7% of what users rate, 5.5× the rate for other agents. Ranks 3rd of 6 on “Good voice conversations”. Users say they call their dot and talk to it while walking, at the gym or picking up food. They describe the conversation as more fluid than typing, and some rate it above Grok Bot's voice.
  Evidence:
  - 2026-10-08, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i’ve been using openai’s dots, and voice makes a big difference. talking with my dot feels more fluid than typing or sending voice notes. i can think out loud and keep the conversation going. it makes everyday tasks feel easier.” [source](https://twitter.com/1740075805538230272/status/2108054140240973964)
  - 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “grok bot's voice conversation experience isn't even close to what chatgpt dots offers. that said, i am not giving up on grok bot just yet. it has potential.” [source](https://twitter.com/723754228238438400/status/2111523723983638803)
- **Its usage allowance goes further** ([Usage allowance lasts](https://feedbackbench.com/criteria/billing.usage_cap.md)). 1.5×: 6% of what users rate, 1.5× the rate for other agents. Ranks 2nd of 6 on “Usage allowance lasts”. Users say work the dot does itself does not count against their weekly limits, so they run it all day. Some send work to it when other agents hit their rate limits.
  Evidence:
  - 2026-10-11, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “you guys should not be sleeping on @openai dots. insane that this is allowed within your subscription and won’t deduct from your weekly limits. offload the heavy work that burns your weekly usage like computer use, and thank me later.” [source](https://twitter.com/2105421372738445312/status/2109313246654054875)
  - 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@connortalksai grok bot runs my business, meta muse handles my daily chores, and chatgpt dots is where i go when the other two hit rate limits.” [source](https://twitter.com/1768748968186695680/status/2111656450758164784)

### Why people leave

- **Replies are slow and calls drop** ([Chatting with it](https://feedbackbench.com/criteria/ask.md)). 1.7×: 22% of what users rate, 1.7× the rate for other agents. Ranks 8th of 8 on “Chatting with it”. Posts describe long waits between actions and minutes of thinking on simple questions. Some say a turn stalls until the user prompts it again. Calls hang up when the phone locks or take seconds to connect.
  Evidence:
  - 2026-10-14, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “anybody experiencing extreme degradation on @openai dots? currently mine thinks for 5-10min to reply on the simplest of questions... [link]” [source](https://twitter.com/114670417/status/2110421847372374051)
  - 2026-10-14, The dots demo, take two | OpenAI DevDay 2026 (YouTube): “when i call my dot, if my phone is locked, it hang up.. it would be great if the call can continue in the background.” [source](https://www.youtube.com/watch?v=fHEIw5CcN5U&lc=Ugz9Nm9LoRZTJkkGU4h4AaABAg)
- **Tasks get handed off or left undone** ([Getting tasks done](https://feedbackbench.com/criteria/act.md)). 9th: of 9 on “Getting tasks done”. 28% of what users rate. Users say the dot passes instructions to Codex instead of doing the work itself, and often does it poorly. Posts describe an agent that does not know what it is connected to, needs micromanaging, and sometimes does nothing at all.
  Evidence:
  - 2026-10-14, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “its totally asinine that @chatgpt dot just dispatches instructions to codex, usually poorly. @bot just works” [source](https://twitter.com/1706492788575997952/status/2110212202737401870)
  - 2026-10-13, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i let @openai dot work overnight on a project as i had 40% usage and expected a reset. dot did nothing all night for nine hours. asi is here. [link]” [source](https://twitter.com/2058615230263963648/status/2109973361413554327)
- **Limits arrive sooner than users expect** ([Price and value](https://feedbackbench.com/criteria/billing.md)). 5th: of 7 on “Price and value”. 15% of what users rate. Users say the dot now opens Codex sessions even for small tasks. That uses up the quota they thought it saved. Heavy users report hitting an abuse prevention limit that pauses the agent.
  Evidence:
  - 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@1edgar @openai dot now habitually uses codex, so the early method that did not consume quota has almost disappeared. for a very small task, it will open a codex session.” [source](https://twitter.com/2075574700793851904/status/2111387347308970062)
  - 2026-10-13, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “chatgpt dot: "your dot is on a break. congrats you are in the very top users and have hit our abuse prevention limit. check back in a bit!" dang it....” [source](https://twitter.com/1588164572824801282/status/2110071120770843047)

### Head to head

- **ChatGPT Dots vs Grok Bot**: ChatGPT Dots ahead in 21 of 64 posts that pick a side, 34 even. Grok Bot leads 43 to 21 across 98 posts, with 34 even. Users call Grok Bot more developed, snappier and cheaper at the lower tiers. Dots wins on usage that lasts. In Grok Bot's own subreddit, Dots leads 4 to 3.
  - 2026-10-07, Grok Bot Now Uses Claude Opus 5.5 + More (MASSIVE Update) (YouTube): “i thought i would like dot better and looks very cute but grok bot is more developed and cheaper at the lower tiers.” [source](https://www.youtube.com/watch?v=Z8H3MvLJwfs&lc=UgzWy4FMgkHhEcdU-aF4AaABAg.AbezHQdGexuAbfGBNZwo5r)
  - 2026-10-14, I Made Grok Bot, Dots & Muse Do the Same 3 Jobs (RAW RESULTS) (YouTube): “i cancelled grok after like an hour when like 4 prompts used up my creds and did absolutely nothing! while i use dots none stop all day with like no usage. it does much more i think dots wins by far” [source](https://www.youtube.com/watch?v=fbm5LevEJyI&lc=UgxEH_-OxMQwuHirZad4AaABAg)
- **ChatGPT Dots vs Muse**: ChatGPT Dots ahead in 12 of 53 posts that pick a side, 52 even. Muse leads 41 to 12 across 105 posts, with 52 even. Users say Muse is faster and handles tasks and plugins where Dots breaks. Some users pick Dots because they distrust Meta's data handling.
  - 2026-10-10, 메일·투자·PPT·쇼핑까지 맡겨봤습니다｜ChatGPT dots 솔직 리뷰 (YouTube): “ah! and if i get a refund for the subscription, they said the dot menu will completely disappear, so i transferred all the tasks i did with the dot to muse and told them to do the automated tasks as they were, and they did well... even the speed is faster with muse...” [source](https://www.youtube.com/watch?v=2g0mSyLfdwg&lc=Ugx4sRm83RxCTdOga-B4AaABAg)
  - 2026-10-01, ChatGPT Dots Is Here! (Is This OpenAI's Grok Bot Killer?) (YouTube): “regarding the work use case, i tend to disagree. i've connected my dot to a isolated password vault and i was able to have it book tickets as well and other useful things consumer wise. given meta's data handling and use policies, i just cant trust muse.” [source](https://www.youtube.com/watch?v=z5X1eMU6isI&lc=UgzCb1kImDMj1_SfRTJ4AaABAg)
- **ChatGPT Dots vs OpenClaw**: ChatGPT Dots ahead in 21 of 34 posts that pick a side, 15 even. Dots leads 21 to 13 across 49 posts, with 15 even. Users say Dots sets up in minutes instead of hours. OpenClaw wins on memory, local models and having no usage limits.
  - 2026-10-09, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i ditched openclaw for openai's dots, and setup took minutes instead of hours [link] [link]” [source](https://twitter.com/65119912/status/2108361865566806246)
  - 2026-10-12, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai dots memory is atrocious i take openclaw active memory for granted i guess but this is chatbot level of context only memory but dot has multiple issues i dislike overall” [source](https://twitter.com/2079171205417095168/status/2109787068654342442)

### What would move its score

Rule: For each area where an agent's customer love is below the category norm: the Feedback Score and rank it would have if that one area were at the norm (0.5) and everything else stayed the same, computed with the same formula as the ranking. An area's pull follows its weight, so an area many users of the category rate moves the score more than one they rarely mention.

| Area at the category norm | Feedback Score | Change | Position |
|---|---|---|---|
| Getting tasks done | 42.1 | +2.9 | 8th of 9 |
| Talking to the agent | 40.5 | +1.3 | 8th of 9 |
| Running on its own | 40.2 | +1.0 | 8th of 9 |
| Staying in control | 40.2 | +0.9 | 8th of 9 |
| Connecting accounts, sites and devices | 39.8 | +0.5 | 8th of 9 |
| Dealing with the vendor | 39.7 | +0.5 | 8th of 9 |
| Getting access and set up | 39.7 | +0.5 | 8th of 9 |
| Paying for the agent | 39.6 | +0.4 | 8th of 9 |
| Trusting it with data and safety | 39.4 | +0.1 | 8th of 9 |

Getting tasks done would move the score most. At the category norm, it lifts the Feedback Score from 39.2 to 42.0, and Dots still ranks 8th. Chatting with it would add 1.3 points, and no other area adds more than 1.0. No single area lifts Dots out of 8th place.

### Six weeks

Users first posted in the week of 28 September, when 6,585 did. A release in ChatGPT came on 29 September. Dots was not ranked that week. Users then fell to 3,158 (12% share of voice, 7th) and then to 1,914 (9%, 8th).

## Top requests

What users ask to add or change, most asked first. 1020 user-weeks ask for something. Requests do not change the Feedback Score. Rule: A separate pass by a large language model reads every counted post and extracts what the user asks the agent or its vendor to add or change, with the criteria it maps to and a short normalised wording; it does not touch the labels or the Feedback Score. A large language model groups the wordings within each criterion (the first criterion the request maps to) into themes, and a second pass merges themes that ask for the same change under different criteria; a merged theme sits under the criterion most of its requests map to, and code counts it once. A theme counts distinct user-weeks that ask for it, per agent; across agents, one user-week per agent. Themes asked in fewer than 2 user-weeks, and requests that share no theme, are not shown. Examples: up to 3 posts per theme from different users, without slurs: whole posts of 60 to 450 characters with at least 8 words beyond handles and links, posts of at most 280 characters first, then the newest.

| Rank | Request | Criterion | User-weeks | Posts |
|---|---|---|---|---|
| 1 | Availability in the EU and Europe | [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | 75 | 79 |
| 2 | Availability on the Plus plan | [Free tier, trial or required plan to start](https://feedbackbench.com/criteria/billing.entry_plan.md) | 70 | 71 |
| 3 | Multiple agents per account | [Multiple agents, teams and delegation between them](https://feedbackbench.com/criteria/auto.multi_agent.md) | 31 | 32 |
| 4 | Availability on cheaper plans than top tier | [Free tier, trial or required plan to start](https://feedbackbench.com/criteria/billing.entry_plan.md) | 22 | 23 |
| 5 | Availability in the UK | [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | 18 | 18 |
| 6 | Standalone dedicated app, not embedded | [Apps offered for phone, tablet and desktop](https://feedbackbench.com/criteria/access.platform.md) | 13 | 13 |
| 7 | Connect multiple computers to one agent | [Access to the user's own phone or computer](https://feedbackbench.com/criteria/connect.local_device.md) | 12 | 12 |
| 8 | Instant call pickup without fake ringing | [Response and task completion speed](https://feedbackbench.com/criteria/ask.speed.md) | 12 | 12 |
| 9 | Selectable and custom agent voices | [Speaking with the agent by voice or call](https://feedbackbench.com/criteria/ask.voice.md) | 12 | 12 |
| 10 | Global availability beyond launch countries | [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | 11 | 11 |
| 11 | Home and lock screen widgets | [Apps offered for phone, tablet and desktop](https://feedbackbench.com/criteria/access.platform.md) | 11 | 11 |
| 12 | Pro plan availability in Europe and UK | [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | 10 | 10 |

### 1. Availability in the EU and Europe

- ChatGPT Dots, 2026-10-17, Put your dot on call (YouTube): “would love to, but at least im paying for that in switzerland even it is not enabled 😂” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgwXSby0j1eVg0b-0sF4AaABAg)
- ChatGPT Dots, 2026-10-16, Put your dot on call (YouTube): “i hope your dot checks these yt comments for feedback as well, please make it work in the eu and support more chat apps (like google chat integration)” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgxzokKTeH7A74Gg8o14AaABAg)
- ChatGPT Dots, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@robinebers and chatgpt dots aren't available in the eu yet, last i checked. astra 6.1 is going to need to be a model that competes with opus 5.5, currently anyway, with good usage limits, in order for them to start bouncing back.” [source](https://twitter.com/1905900669678039040/status/2111148661694746666)

### 2. Availability on the Plus plan

- ChatGPT Dots, 2026-10-18, Put your dot on call (YouTube): “make it available for plus users.......what's wrong with you? we pay...so give us this service please” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgytbHLaCyZ_SvHw_114AaABAg)
- ChatGPT Dots, 2026-10-17, Put your dot on call (YouTube): “yeah, this is such bullshit that dots hasn't rolled out to paid plus users. it really makes me want to jump ship.” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=Ugy8q4Kc2WByrCqXfL94AaABAg)
- ChatGPT Dots, 2026-10-17, Put your dot on call (YouTube): “i need dots in my plus acount, please, i'm an autonomous worker, just one orchestra worker and i think as insurance broker and real state could be great that you help me, three years working with you already and top fan! ❤” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgxxCyWE5m_TW_mCeRx4AaABAg)

### 3. Multiple agents per account

- ChatGPT Dots, 2026-10-17, Put your dot on call (YouTube): “1) when will it be dots instead of 'dot' (singular) 2) when can it be able to call people on the phone on my behalf? 3) when can it 100% automate my computer? (without me having to resort to codex) these are the only things i'm interested in.” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgwtJpQTH1b2AcFIiYJ4AaABAg)
- ChatGPT Dots, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “why are openai’s dots called dots if you can only have one? i need more.” [source](https://twitter.com/780497815818625028/status/2111579561267999121)
- ChatGPT Dots, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “my @openai "dot" only has 1 purpose. tell me when "dot(s)" is released. otherwise, it's ignored entirely. @openaidevs @thsottiaux single dot is useless for my use cases. [link]” [source](https://twitter.com/733419117139496960/status/2111552769383620779)

### 4. Availability on cheaper plans than top tier

- ChatGPT Dots, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “the fact that grokbot @bot and meta's muse @muse are available for most low-tier subscribers and even for free should be a wake up call for openai's dots @openai. like i wanna use dots but the fact that you have to pay $100/mo for it is nuts” [source](https://twitter.com/1944393310690263040/status/2111158600760467804)
- ChatGPT Dots, 2026-10-10, Set up your dot in the ChatGPT mobile app (YouTube): “@mspardal123indeed, and i use it too. but i’ve been using chat since dec 23’. we go wayyy back😅 i mean give us a dot with a limit and then if it’s worth it maybe we’ll join the next tier.” [source](https://www.youtube.com/watch?v=p5KxBTUncoE&lc=Ugz554lNTjDeusOo_eh4AaABAg.AbklmO485BkAbmjohLCJ3E)
- ChatGPT Dots, 2026-10-10, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i want to try using chatgpt "dots"! but it costs $100... i'm a bit hesitant. is everyone using pro? [link]” [source](https://twitter.com/1940904794/status/2108938116111458401)

### 5. Availability in the UK

- ChatGPT Dots, 2026-10-17, Put your dot on call (YouTube): “i recognise that chatgpt is impressive, and i am unsure why you have not introduced dote in the uk or why versions vary across countries ?” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=Ugyd6-lb75J-ONzI2xR4AaABAg)
- ChatGPT Dots, 2026-10-09, Set up your dot in the ChatGPT mobile app (YouTube): “i can’t access this because i’m in the uk. not sure why it’s not available to us” [source](https://www.youtube.com/watch?v=p5KxBTUncoE&lc=UgzvwbQfo0eeFwVNzzV4AaABAg)
- ChatGPT Dots, 2026-10-04, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “uk users having to wait for access to @meta's muse or @openai's dots is a prime example of the lack of ai strength in this country. we need to do better, we can do better.” [source](https://twitter.com/25303081/status/2106812615574114541)

### 6. Standalone dedicated app, not embedded

- ChatGPT Dots, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai dot is the first assistant/companion i’ve tried, and it’s pretty good. it definitely helps me conserve my decision budget. i just wish it was a separate app. when i open chatgpt on ios or codex on macos, i have a different job in mind than checking what my dot is up to.” [source](https://twitter.com/612373/status/2111527996322689129)
- ChatGPT Dots, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@bot(plus @x) > @muse > @openai dot dot needs a standalone app! it makes a huge difference” [source](https://twitter.com/3181743769/status/2110834369204236585)
- ChatGPT Dots, 2026-10-02, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “hot take: chatgpt dots should have their own app. feels odd just being a sidebar tab in the corner. it's a very distinct thing” [source](https://twitter.com/1042799721948098560/status/2106127406750716025)

### 7. Connect multiple computers to one agent

- ChatGPT Dots, 2026-10-16, Put your dot on call (YouTube): “can you please allow it have access to multiple local computers in parallel? seems simple to implement but can boost productivity massively.” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgzfpLILq_r2YELGKnl4AaABAg)
- ChatGPT Dots, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “it might be a niche thing, but i wish a chatgpt dot could connect to multiple always-on desktop computers, to remote control them from each other.” [source](https://twitter.com/14751214/status/2111234147830415472)
- ChatGPT Dots, 2026-10-04, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@thsottiaux request for device connection for chatgpt dots i want it to be able to connect to multiple macs!!” [source](https://twitter.com/3830241/status/2106574328070607218)

### 8. Instant call pickup without fake ringing

- ChatGPT Dots, 2026-10-16, Put your dot on call (YouTube): “it would be amazing if calls didn't take so long to connect or dots didn't suddenly lose access to your codex all the time. when integration isn't an issue, dots are really powerful” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgzzwlgbTYDmrMUlTQ14AaABAg)
- ChatGPT Dots, 2026-10-10, ChatGPT Dots Is WAY More Powerful Than You Think (YouTube): “i wish that the call didn’t ring 3 or 4 times. i was used to the voice mode that starts immediately.” [source](https://www.youtube.com/watch?v=WXhOxfPECnM&lc=UgwfZspntayrj7ZAmnZ4AaABAg)
- ChatGPT Dots, 2026-10-09, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “my @chatgpt dot “call” rings 5x then fails my @grok @bot picks up within 1 sec and gets to work the fake phone call set up is a terrible product decision, especially when your “always on assistant” just… isn’t on? [link]” [source](https://twitter.com/1756394384655003648/status/2108566954474938847)

### 9. Selectable and custom agent voices

- ChatGPT Dots, 2026-10-16, Put your dot on call (YouTube): “please give us the ability to change our dots voices separately from a chat gpt voices thank you” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgyC7e5AW_XrDTdr6ud4AaABAg)
- ChatGPT Dots, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@openai dots should support female voices😀, the male voice lacks clarity.” [source](https://twitter.com/2110862543891976192/status/2110938927951720701)
- ChatGPT Dots, 2026-10-06, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i love my @openai dots. but i’d love it more if he was old and british. how do i change my voice 😭🤞 [link]” [source](https://twitter.com/1546805400678600710/status/2107276987140370450)

### 10. Global availability beyond launch countries

- ChatGPT Dots, 2026-10-13, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@kanishkanarayan too many restrictions, we cant get openai's dot, no chance of meta's muse neither, which means the public cant keep up with usa publics innovation.” [source](https://twitter.com/1601993174548971520/status/2109936858826445125)
- ChatGPT Dots, 2026-10-09, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@alexandr_wang muse is the goat. one month in and the whole field woke up, openai dots, manus cue, grokbot …etc, full robot war. beautiful. quick question tho, the other 190 countries were also invited, or is “global” still a concept?🤣” [source](https://twitter.com/1239214643219841024/status/2108500192119140745)
- ChatGPT Dots, 2026-10-02, ChatGPT Dots Is Insane... But Gemini’s NEW Argon Is EVEN Bigger (YouTube): “to bad i am from `european economic area` so gotta wait till they rollout dots worldwide :p” [source](https://www.youtube.com/watch?v=kmo9Mk_gIUk&lc=Ugy-bZF6ZmWLBtM8I-p4AaABAg)

### 11. Home and lock screen widgets

- ChatGPT Dots, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “a little ux wishlist for chatgpt dots, @openai 👀 • bring back swipe-right-to-reply • add a search bar for our dot conversations • give dots an android home-screen widget small touches that would make everyday use so much nicer. especially that widget 🥹” [source](https://twitter.com/1843728766541148160/status/2111313894115520909)
- ChatGPT Dots, 2026-10-13, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@reach_vb can you all make a chatgpt dot shortcut or widget please i'm sick of double tapping into the app then my fat booty dot. its losing me 100ks of revenue as we speak wasting my time” [source](https://twitter.com/766925674909052928/status/2110034298879496248)
- ChatGPT Dots, 2026-10-10, Set up your dot in the ChatGPT mobile app (YouTube): “rrally need seperate app or widget for phones. i dnt wanna keep going through the chatgpt app. do same as muse” [source](https://www.youtube.com/watch?v=p5KxBTUncoE&lc=UgzoNldBZ9lQtEZdAMF4AaABAg)

### 12. Pro plan availability in Europe and UK

- ChatGPT Dots, 2026-10-05, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i can’t use openai dots on my pro plan because i’m in the eu. so i’m looking at grok bot 👀 [link]” [source](https://twitter.com/1956392094479646728/status/2107179470708941189)
- ChatGPT Dots, 2026-10-05, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “📰 openai dots: agents continue to work, not yet for eu-pro openai launched dots, agents with their own cloud computer that also work when you don't. what they can do… 👉 read more: [link] #ai #artificialintelligence #ainews #openai #dots #chatgpt [link]” [source](https://twitter.com/854974516795236352/status/2107000591071293829)
- ChatGPT Dots, 2026-10-03, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai's dots launched this week. if you're a pro user in europe, you're not invited. again. always-on agents are redefining how work gets done, and we're watching from the waiting list. the future doesn't wait for brussels to finish its paperwork.” [source](https://twitter.com/132249056/status/2106299039331320130)

## Ranks highest on

Speaking with the agent by voice or call (3rd of 6)

## Ranks lowest on

Access to the user's own phone or computer (5th of 5), Response and task completion speed (7th of 7), Where the agent runs: its own computer and always-on running (6th of 6)

Position on a criterion: its place among the agents with at least 30 rated user-weeks there, ordered by customer love. It ranks high when both its position and its 95% interval put it in the top half, low when both put it in the bottom half; neither otherwise, or when fewer than 4 agents qualify.

## All 53 criteria

Criterion love: 0.5 is the category norm. n: rated user-weeks.

### Paying for the agent: 5th of 7 (customer love 0.454, n 482)

Users praise the allowance, where it ranks 2nd, but call the price unfair for what it does, last of 6. Posts describe Codex sessions and abuse limits eating into plans. Users ask for it on the Plus plan and on cheaper plans.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Usage allowance size, exhaustion and reset wait](https://feedbackbench.com/criteria/billing.usage_cap.md) | 2nd of 6 | 0.564 | 0.526–0.602 | 202 | 67 | 135 |
| [Subscription price versus value received](https://feedbackbench.com/criteria/billing.price_level.md) | 6th of 6 | 0.394 | 0.367–0.422 | 126 | 30 | 96 |
| [Agent spends usage on overhead and costly choices](https://feedbackbench.com/criteria/billing.token_efficiency.md) | 3rd of 5 | 0.514 | 0.484–0.543 | 85 | 30 | 55 |
| [Free tier, trial or required plan to start](https://feedbackbench.com/criteria/billing.entry_plan.md) | 5th of 6 | 0.403 | 0.376–0.432 | 82 | 21 | 61 |
| [Plan contents, bundles and usage meters unclear](https://feedbackbench.com/criteria/billing.plan_clarity.md) | Too few posts | 0.488 | 0.471–0.515 | 24 | 1 | 23 |
| [Buying, upgrading, cancelling and refunding go wrong](https://feedbackbench.com/criteria/billing.purchase_flow.md) | Too few posts | 0.491 | 0.485–0.496 | 9 | 0 | 9 |

Most recent posts:

- Praise, 2026-10-16, OpenAI Dots Is Already Changing How AI Agents Look (YouTube): “i think people are sleeping on these dot agents. i've completely redone many workflows and processes to take advantage of the dots (and the 100% free usage for a month).” [source](https://www.youtube.com/watch?v=Sj9SfzUXTjM&lc=Ugx6QffdN7T1jH9f9AB4AaABAg)
- Praise, 2026-10-14, I Made Grok Bot, Dots & Muse Do the Same 3 Jobs (RAW RESULTS) (YouTube): “i cancelled grok after like an hour when like 4 prompts used up my creds and did absolutely nothing! while i use dots none stop all day with like no usage. it does much more i think dots wins by far” [source](https://www.youtube.com/watch?v=fbm5LevEJyI&lc=UgxEH_-OxMQwuHirZad4AaABAg)
- Praise, 2026-10-14, OpenAI Co-Founder: Start Building With AI Before You Feel Ready | Greg Brockman (YouTube): “chatgpt dots changed the way i work daily. what a bargain. still, security awareness and blast radius should always be set up” [source](https://www.youtube.com/watch?v=qy8Gr27yLMk&lc=Ugxx45cbwsoueyGfQ_l4AaABAg)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@1edgar @openai dot now habitually uses codex, so the early method that did not consume quota has almost disappeared. for a very small task, it will open a codex session.” [source](https://twitter.com/2075574700793851904/status/2111387347308970062)
- Complaint, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “is there anyone who is completely working with glock @bot and @chatgpt dots? with the current usage restrictions, i can only manage to organize emails satisfactorily, and if i try to post on x every day, it won't last a week.” [source](https://twitter.com/267007697/status/2110728861629723106)
- Complaint, 2026-10-14, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i was using chatgpt dot a lot because it was free, but i got stopped. [link]” [source](https://twitter.com/985914910813995010/status/2110286114187804784)

### Getting access and set up: 6th of 7 (customer love 0.414, n 277)

Setup itself is close to the category norm, and some users say it took minutes. The complaints are about who can get it. Users ask for it in the EU and the UK, on the Plus plan and on cheaper plans.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | 3rd of 4 | 0.458 | 0.414–0.500 | 158 | 12 | 146 |
| [Effort needed to set up and train the agent](https://feedbackbench.com/criteria/access.setup.md) | 4th of 6 | 0.505 | 0.482–0.529 | 62 | 31 | 31 |
| [Apps offered for phone, tablet and desktop](https://feedbackbench.com/criteria/access.platform.md) | 4th of 5 | 0.473 | 0.452–0.493 | 42 | 6 | 36 |
| [Account creation, sign-in and identity checks fail](https://feedbackbench.com/criteria/access.sign_in.md) | Too few posts | 0.480 | 0.471–0.488 | 20 | 0 | 20 |

Most recent posts:

- Praise, 2026-10-12, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@jayair this is literally just openai's dots in codex on both a mac mini and macbook. very simple setup and works right now. don't want what you can already have!” [source](https://twitter.com/237428719/status/2109457659221229850)
- Praise, 2026-10-10, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “create your chatgpt dot from your phone — ios and android, right in the chatgpt app. i used to wait until i was back at the laptop to spin one up. now i can start a dot on the train and have it ready when i sit down. nice one @thsottiaux [link]” [source](https://twitter.com/1894226409356496903/status/2108876227805061243)
- Praise, 2026-10-10, ChatGPT Dots a Spaces v praxi: Jak to sakra funguje? Furt to nechápu... (YouTube): “dots works for me for some unknown reason even without a vpn on a czech phone in the czech republic, and i haven't left the eu in the last 5 years. but i am excited about it and it works really great for me.” [source](https://www.youtube.com/watch?v=2Z2FGDTyuUU&lc=UgwESG264m4-c3ubjAF4AaABAg)
- Complaint, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@bot(plus @x) > @muse > @openai dot dot needs a standalone app! it makes a huge difference” [source](https://twitter.com/3181743769/status/2110834369204236585)
- Complaint, 2026-10-15, OpenAI Dots: ИИ, которому больше не нужно все объяснять? Новый способ работать с AI (YouTube): “i am in bulgaria. the dot appeared, gave it the first task, and it disappeared, saying that this region is restricted, while emails come with an offer to try it.” [source](https://www.youtube.com/watch?v=gs2oOsi18Pg&lc=Ugxm4nw3nu5VYREBHJp4AaABAg)
- Complaint, 2026-10-12, We Tested ChatGPT Dots So You Don't Waste $100 (YouTube): “as far as i understand you are correct. i haven't used grok bot but i've setup dots and that part is frustrating. but i'm certain openai will drop more updates soon” [source](https://www.youtube.com/watch?v=SZMtfvLvUM0&lc=UgyRZeO1lhqetly5Xbl4AaABAg.AbkesbZTs0UAbqYYw4AEpx)

### Connecting accounts, sites and devices: 7th of 7 (customer love 0.391, n 362)

Users say linked accounts drop their connection, and reaching their own phones and computers ranks last of 5. Posts describe plugins that stop working. Users ask to connect more than one computer to one agent.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Which services have a connector or integration](https://feedbackbench.com/criteria/connect.coverage.md) | 6th of 7 | 0.454 | 0.421–0.485 | 125 | 66 | 59 |
| [Linking accounts and staying logged in](https://feedbackbench.com/criteria/connect.account_link.md) | 6th of 6 | 0.432 | 0.409–0.459 | 87 | 4 | 83 |
| [Access to the user's own phone or computer](https://feedbackbench.com/criteria/connect.local_device.md) | 5th of 5 | 0.401 | 0.368–0.432 | 84 | 23 | 61 |
| [Websites block the agent with captchas or IP bans](https://feedbackbench.com/criteria/connect.site_blocking.md) | 6th of 7 | 0.497 | 0.469–0.526 | 51 | 10 | 41 |
| [Moving files between agent, apps and user](https://feedbackbench.com/criteria/connect.file_transfer.md) | 2nd of 3 | 0.485 | 0.465–0.507 | 32 | 6 | 26 |

Most recent posts:

- Praise, 2026-10-11, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@kylespatel i’ve been very impressed with the new dot from openai and i really like the apple text app plugin in the desktop app. now my dot can send and reply to my text messages. maybe soon we can get it to take and make phone calls too. :)” [source](https://twitter.com/2094174287246524416/status/2109211165620596790)
- Praise, 2026-10-08, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@lukenneth_ @jankeesvw @t3dotcodes @theo i thought they enabled that in t3code? either way i use my chatgpt dot with the new t3code mcp to connect and manage all my t3code projects. it orchestrates them all.” [source](https://twitter.com/1965828809602768897/status/2108247438536651198)
- Praise, 2026-10-07, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i just called chungus (my openai dot) on the phone while picking up mexican food and asked it to transfer some rex benchmarks from last night on my mac studio to my macbook pro via tailscale and i just got to my desk and there they are. what a fucking time to be alive.” [source](https://twitter.com/12819682/status/2107907073992450112)
- Complaint, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “anyone else getting this with openai dots when trying to read a thread started in codex? “i tried that thread id, but it returned a thread-format error before i could read it. i’ll check whether there’s another supported way to open it.”” [source](https://twitter.com/1252363544/status/2111813876182597867)
- Complaint, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@chefmcmakin haven’t tried grok bot yet. still using chatgpt dots but seems like a lot of sites block it.” [source](https://twitter.com/6484412/status/2111869526879129657)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “having some issues with @chatgpt dots browser connectivity. anybody have any suggestions for a smoother experience?” [source](https://twitter.com/410410934/status/2111311020207829207)

### Talking to the agent: 8th of 8 (customer love 0.346, n 734)

Users talk about chatting 1.7 times as often as for other agents. Voice calls draw praise and rank 3rd. Speed ranks last, and users say threads do not stay organised.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Speaking with the agent by voice or call](https://feedbackbench.com/criteria/ask.voice.md) | 3rd of 6 | 0.549 | 0.514–0.587 | 235 | 103 | 132 |
| [Response and task completion speed](https://feedbackbench.com/criteria/ask.speed.md) | 7th of 7 | 0.300 | 0.276–0.326 | 233 | 31 | 202 |
| [Threads, chats and parallel tasks organisation](https://feedbackbench.com/criteria/ask.conversation.md) | 5th of 5 | 0.451 | 0.417–0.483 | 153 | 45 | 108 |
| [App screens, layout, rendering, localisation and stability](https://feedbackbench.com/criteria/ask.app_client.md) | 5th of 6 | 0.458 | 0.430–0.487 | 88 | 14 | 74 |
| [Avatar, personality and tone of replies](https://feedbackbench.com/criteria/ask.persona.md) | 5th of 6 | 0.431 | 0.396–0.471 | 83 | 50 | 33 |
| [Reaching the agent through messaging apps](https://feedbackbench.com/criteria/ask.channel.md) | 4th of 5 | 0.469 | 0.441–0.496 | 39 | 20 | 19 |

Most recent posts:

- Praise, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “grok bot's voice conversation experience isn't even close to what chatgpt dots offers. that said, i am not giving up on grok bot just yet. it has potential.” [source](https://twitter.com/723754228238438400/status/2111523723983638803)
- Praise, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “conversation with chatgpt dots on gpt-live-1.1 is super useful.” [source](https://twitter.com/1298232496866226180/status/2110937502202941515)
- Praise, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i didn’t expect openai’s dots to be good with kids but it was joking with her and making her laugh. [link]” [source](https://twitter.com/5550612/status/2110535512469672213)
- Complaint, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i have add swipe reply and still have swipe to back to last screen. just like message app. but @chatgpt dots cannot do reply at all now。。。” [source](https://twitter.com/1003913730/status/2111655368850116947)
- Complaint, 2026-10-18, Put your dot on call (YouTube): “my dots really busy with other clients so he takes up to 30 seconds” [source](https://www.youtube.com/watch?v=GBOZIRK1ujw&lc=UgxeVJSFqL7yPsmIOw54AaABAg.Ac2DCbfx8hcAc5C7RNj5Jh)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “ngl the ui ux for @openai dot is shockingly bad and they are lazy af” [source](https://twitter.com/16164618/status/2111249310134218890)

### Getting tasks done: 9th of 9 (customer love 0.344, n 939)

This is its largest area. Users say it stops short of finishing, misses instructions and misreports what it did, and it ranks last on all three. Building code and research hold up better. Posts describe work handed off to Codex.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Stops short, or hands the task back](https://feedbackbench.com/criteria/act.stops_short.md) | 6th of 6 | 0.374 | 0.346–0.405 | 151 | 40 | 111 |
| [Writing code and building apps or websites](https://feedbackbench.com/criteria/act.build.md) | 4th of 6 | 0.492 | 0.439–0.548 | 136 | 122 | 14 |
| [Web research, monitoring and summarising information](https://feedbackbench.com/criteria/act.research.md) | 3rd of 7 | 0.532 | 0.472–0.584 | 109 | 105 | 4 |
| [Wrong facts, hallucinations and careless mistakes](https://feedbackbench.com/criteria/act.accuracy.md) | 6th of 7 | 0.459 | 0.431–0.488 | 108 | 23 | 85 |
| [Following instructions, rules and corrections](https://feedbackbench.com/criteria/act.instructions.md) | 6th of 6 | 0.451 | 0.424–0.479 | 95 | 22 | 73 |
| [Email triage, cleanup and drafting replies](https://feedbackbench.com/criteria/act.email.md) | 6th of 7 | 0.457 | 0.394–0.520 | 90 | 78 | 12 |
| [Creating writing, images, video and social posts](https://feedbackbench.com/criteria/act.content.md) | 5th of 6 | 0.488 | 0.452–0.527 | 87 | 71 | 16 |
| [Operating a browser or desktop by clicking](https://feedbackbench.com/criteria/act.computer_use.md) | 7th of 7 | 0.432 | 0.401–0.461 | 73 | 26 | 47 |
| [Forms, applications, claims and life admin paperwork](https://feedbackbench.com/criteria/act.admin.md) | 5th of 7 | 0.475 | 0.393–0.559 | 64 | 61 | 3 |
| [Says it did something it did not do](https://feedbackbench.com/criteria/act.false_reports.md) | 5th of 5 | 0.446 | 0.433–0.459 | 58 | 0 | 58 |
| [Booking appointments, reservations, travel and tickets](https://feedbackbench.com/criteria/act.reservations.md) | 5th of 5 | 0.434 | 0.380–0.480 | 33 | 22 | 11 |
| [Contacting other people and businesses for you](https://feedbackbench.com/criteria/act.outreach.md) | 5th of 6 | 0.477 | 0.442–0.511 | 33 | 25 | 8 |
| [Calendar entries, conflicts and schedule changes](https://feedbackbench.com/criteria/act.calendar.md) | Too few posts | 0.487 | 0.436–0.524 | 21 | 19 | 2 |
| [Buying products, orders and purchase judgment](https://feedbackbench.com/criteria/act.shopping.md) | Too few posts | 0.486 | 0.461–0.508 | 19 | 12 | 7 |
| [Budgets, transactions, bills and investments](https://feedbackbench.com/criteria/act.finance.md) | Too few posts | 0.477 | 0.419–0.520 | 18 | 16 | 2 |

Most recent posts:

- Praise, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “was skeptical of the utility of chatgpt dots, but i set up a long-term task for finding me a good deal for a quality mechanical keyboard and it did a pretty good job” [source](https://twitter.com/1508937527364829191/status/2111615602456400292)
- Praise, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@_dylanga i'm going to do a video on this, but i used my openai dot to cancel a zombie gym subscription where i had looked at my credit card statement three statements in a row and didn't do anything, but dashing off a sentence to my dot was sufficient to overcome the inertia.” [source](https://twitter.com/1980768431227228160/status/2111727717683261780)
- Praise, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i was able to operate blender and create 3dcg through chatgpt dots. [link]” [source](https://twitter.com/128910776/status/2111750250289615229)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@bot grok bot is far better at trading than alternatives. really impressed so far. fridays trade on mnq made +8.25r (obviously short) $qqq muse is profitable and @chatgpt dots are atrocious” [source](https://twitter.com/1715904152256053248/status/2111487904228732989)
- Complaint, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@thsottiaux my chatgpt dot is experiencing recurring automation failures: failed to query executor configuration capabilities; cdp operation attach_tab timed out after 20000ms page.enable / emulation errors please reinitialize or repair the affected executor” [source](https://twitter.com/2048959948760199169/status/2110936022871875791)
- Complaint, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@thsottiaux anyone else having issues with chatgpt dots? @openai manual cloud browser takeover works perfectly, but dot's automation keeps failing: • failed to query executor configuration capabilities • cdp attach_tab timed out • page.enable / emulation errors #chatgpt” [source](https://twitter.com/2048959948760199169/status/2110936502880596407)

### Running on its own: 8th of 8 (customer love 0.398, n 700)

Taking useful initiative draws praise and ranks 3rd of 6. Its own computer and agent teams both rank last. Users ask for more than one agent per account.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Where the agent runs: its own computer and always-on running](https://feedbackbench.com/criteria/auto.cloud_env.md) | 6th of 6 | 0.388 | 0.355–0.422 | 234 | 116 | 118 |
| [Unprompted suggestions, insights and check-ins](https://feedbackbench.com/criteria/auto.proactive.md) | 3rd of 6 | 0.512 | 0.467–0.555 | 204 | 172 | 32 |
| [Multiple agents, teams and delegation between them](https://feedbackbench.com/criteria/auto.multi_agent.md) | 5th of 5 | 0.415 | 0.377–0.454 | 148 | 96 | 52 |
| [Remembering the user and context across sessions](https://feedbackbench.com/criteria/auto.memory.md) | 3rd of 5 | 0.456 | 0.427–0.491 | 111 | 59 | 52 |
| [Scheduled recurring tasks and daily briefings run](https://feedbackbench.com/criteria/auto.routines.md) | 5th of 6 | 0.469 | 0.431–0.507 | 44 | 31 | 13 |
| [Reminders and notifications reach the user correctly](https://feedbackbench.com/criteria/auto.reminders.md) | Too few posts | 0.497 | 0.475–0.518 | 21 | 16 | 5 |

Most recent posts:

- Praise, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “neat my chatgpt dot told me this morning about a storm that’d happen during work later in the day. it very much came lmao” [source](https://twitter.com/957461250111410176/status/2111313717027631567)
- Praise, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai dots, even though i didn't communicate about reopening, it started to process the tasks that had accumulated after the regulation was lifted. thank you very much.” [source](https://twitter.com/439811452/status/2111493677554241800)
- Praise, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai's dots gave me a taste of what a good 24/7 agent looks like, and now grok bot is letting have a team of them!” [source](https://twitter.com/1963054973874515968/status/2111092786418397319)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “why is it that none of my chatgpt dots phone call conversations are then being able to be referenced by dot in the text conversations following it?” [source](https://twitter.com/8717262/status/2111567899508842979)
- Complaint, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “chatgpt dots, it's completely useless.. it's subtle! it doesn't act autonomously at all!” [source](https://twitter.com/1830654438287228928/status/2110647908228063569)

### Staying in control: 7th of 7 (customer love 0.321, n 316)

Users complain about how it asks for approval and say it shows little of what it is doing. It ranks last on both. Choosing the model and shaping it to how they work rank 4th of 5.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Asks before acting: too little or too much](https://feedbackbench.com/criteria/control.approval.md) | 6th of 6 | 0.372 | 0.342–0.403 | 126 | 29 | 97 |
| [Seeing what the agent is doing and has done](https://feedbackbench.com/criteria/control.visibility.md) | 5th of 5 | 0.427 | 0.399–0.455 | 82 | 19 | 63 |
| [Which AI model runs, how good it is, and choosing it](https://feedbackbench.com/criteria/control.model.md) | 4th of 5 | 0.466 | 0.439–0.493 | 75 | 31 | 44 |
| [Configuring behaviour, skills and portability](https://feedbackbench.com/criteria/control.customize.md) | 4th of 5 | 0.416 | 0.387–0.446 | 60 | 21 | 39 |

Most recent posts:

- Praise, 2026-10-16, Muse vs dots vs Grok Bot: The FREE One Almost Won (8 Real Tests) (YouTube): “once. on the pothole report, grok bot uploaded the photo to the city's website without asking first. dots and muse both asked before uploading. all three still stopped before every submit, pay or send button, which was the line i cared about most.” [source](https://www.youtube.com/watch?v=uqOMexa0rzI&lc=UgzXy0L2Y0Y7vkTT2nl4AaABAg.Ac1J7L0zAo6Ac1JdFk2vFT)
- Praise, 2026-10-07, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “yesterday i asked my @chatgpt dots to stop making a document for every draft. too much to keep updated. now one daily linear task holds the drafts and edits, with published links or a rejection kept alongside each one so i can see what still needs my review.” [source](https://twitter.com/2549328379/status/2107697994749640949)
- Praise, 2026-10-06, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i had our dot do a visually understandable status report for the dot's vm that i saw on x. this helped identify the issues that were causing it to stop, and it clearly shows how much it is actually working, which is good! [link]” [source](https://twitter.com/1284128751102472192/status/2107358997829542089)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “anyone else has problems with their chatgpt dot asking for approval for checking mail, even readonly, every day? because apparently the approval expires?” [source](https://twitter.com/13440002/status/2111469461106831725)
- Complaint, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@chatgpt dot is useless for any real work right now because it stops to ask for permission for just about anything i ask it to do. attempting to get it to configure my network and create and assign new vlans and it stops at every single stop to get permission” [source](https://twitter.com/1594179414753001472/status/2111596816999526719)
- Complaint, 2026-10-16, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@reach_vb please tell me you are announcing that auto-review will be removed! it’s so overboard and my dot requires constant handholding, even with custom rules (which themselves get auto-review blocked). i’m sick of authorizing the dumbest things!! openai dots=babysitting” [source](https://twitter.com/1703423105257848832/status/2111011128952103029)

### Trusting it with data and safety: 5th of 6 (customer love 0.432, n 112)

Users rarely raise this area. The complaints are about refusing fair requests, with 0 praise to 33 complaints, and about asking for more access than it needs.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Permission scope requested and isolation between agents](https://feedbackbench.com/criteria/trust.access_scope.md) | 5th of 6 | 0.477 | 0.455–0.500 | 51 | 11 | 40 |
| [Refuses legitimate requests because of guardrails](https://feedbackbench.com/criteria/trust.refusals.md) | 5th of 5 | 0.461 | 0.448–0.473 | 33 | 0 | 33 |
| [How the vendor collects, keeps and uses data](https://feedbackbench.com/criteria/trust.data_use.md) | Too few posts | 0.476 | 0.465–0.485 | 19 | 0 | 19 |
| [Vulnerabilities, prompt injection and cross-user leaks](https://feedbackbench.com/criteria/trust.security.md) | Too few posts | 0.496 | 0.484–0.509 | 12 | 3 | 9 |

Most recent posts:

- Praise, 2026-10-05, OpenAI Dots: İzin almadan bilgisayarımı taramaya başladı. Güvenmeli misiniz? (YouTube): “i trust my dot. i have endured until now, but last week i gave the key to everything, whether it was google workspace, microsoft teams, or personal corporate.” [source](https://www.youtube.com/watch?v=LAO3ztPKsS0&lc=UgxnN9hJhL59xOlxXb14AaABAg)
- Praise, 2026-09-30, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “the best design choice in @openai dots: background proactive research only gets read only access to connected apps. if you ship agents, put that split in the oauth scopes, because a system prompt instruction is a suggestion and a token scope is a wall. [link]” [source](https://twitter.com/2301217708/status/2105180891588792621)
- Praise, 2026-10-11, The $10B AI Assistant Challenging Meta's Muse, Grok's Bots, and OpenAI's Dots (YouTube): “@mark.s.hamilton im very curious how the relationship between the agents will play around this. i am using it with my wife and some other close friends and i do feel the boundaries are set well. i havent push anything sensitive just checking calendar and some stuffs if they like it or not. maybe i should push a little bit harder later and see how it is reacting on behalf each of the "master" lol” [source](https://www.youtube.com/watch?v=Am7IWP8IpEc&lc=UgwBx7S6MU4R6j8lFSZ4AaABAg.AbJTeZzdkOtAbo4pt89-Zr)
- Complaint, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “⚠️attention⚠️ when requesting tasks to chatgpt dots in the local codex, if you do not specify the project folder, it will automatically create a task folder, grant permissions on its own, and start collecting various data on its own!” [source](https://twitter.com/1613121613578530816/status/2110693333937140017)
- Complaint, 2026-10-10, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i've been using openai's dots. although it's a convenient way of interacting with an ai agent as a single point of contact, it refuses to consistently record decisions i make to my local recording system, but it has no issue taking its own notes on me and refuses to export them.” [source](https://twitter.com/1382162150802780164/status/2109013003915948277)
- Complaint, 2026-10-09, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “i like how the first think openai dot does after me creating it with explicitly disabled calendar integration is mentioning events from my calendar to help with them” [source](https://twitter.com/1658458762036891648/status/2108528314994540604)

### Dealing with the vendor: 8th of 8 (customer love 0.380, n 205)

Users raise updates and support less often than for other agents. The complaints are about staying up and responsive, where it ranks last, with 3 praise to 138 complaints.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Outages, errors and agent going unresponsive](https://feedbackbench.com/criteria/vendor.uptime.md) | 7th of 7 | 0.398 | 0.372–0.427 | 141 | 3 | 138 |
| [Got better or worse over time, and update cadence](https://feedbackbench.com/criteria/vendor.updates.md) | 4th of 5 | 0.470 | 0.449–0.493 | 42 | 8 | 34 |
| [Support responses and account bans or lockouts](https://feedbackbench.com/criteria/vendor.support.md) | Too few posts | 0.500 | 0.473–0.536 | 28 | 3 | 25 |

Most recent posts:

- Praise, 2026-10-17, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@steipete @brian_bo_li @chatgpt dots are great already @steipete but i appreciate you continually making them better :)” [source](https://twitter.com/14835779/status/2111511655959933172)
- Praise, 2026-10-05, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai's dot is now in sync on my computer and phone. they weren't kidding about getting the upgrades out asap.” [source](https://twitter.com/4190694130/status/2107093497735962663)
- Praise, 2026-09-30, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “openai dot is amazing... what's amazing is that just yesterday i pointed out the common occurrence of codex falling asleep, and it has already been overturned. an ai that works 24 hours has been born... this is incredible. [link] [link]” [source](https://twitter.com/175553470/status/2105349487535468701)
- Complaint, 2026-10-18, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “chatgpt dots got nerfed today or something because it feels like talking to a person in idiocracy.” [source](https://twitter.com/1243206742122459137/status/2111697812908101754)
- Complaint, 2026-10-18, Les agents IA, ça sert vraiment à quelque chose ? (YouTube): “hello shubham (or your agent haha) indeed, dots is not mature and was released a bit in a rush. in this case, grok bot is definitely more advanced and has more perspective, i think it would be a fairer comparison. have you tested it?” [source](https://www.youtube.com/watch?v=xjK13eSfyTQ&lc=UgxAaWNpvkM2sssVXyZ4AaABAg)
- Complaint, 2026-10-15, X search: ChatGPT dots, ChatGPT dot, OpenAI dots, OpenAI's dot (X): “@diegocabezas01 @openai dots have been nerfed and it's only been like a week. i was a fan but they are so janky and ill conceived when it comes to actually deep workflows that their marketing is essentially a big lie.” [source](https://twitter.com/2098351080308400128/status/2110527023022280794)