On this page
Tibo (@thsottiaux) posted on X that OpenAI is working on expanding the family of frontier models that power ChatGPT dots. The post names GPT-6 Astra and GPT-6.1 Terra, plus others in the future. The account bio on the post reads Codex & ChatGPT @OpenAI.
The main post went up at 2026-11-02 15:50:08 UTC, which is 2026-11-02 23:50:08 Asia/Shanghai. Opening with a short nod to fellow resetters, the text says the team is expanding the frontier-model family behind dots, including Astra, Terra, and more later. It then frames Terra as bringing speed and a nice personality upgrade, and as helping bring dots to more people.
A follow-up in the same thread, posted at 2026-11-02 15:51:03 UTC (23:51:03 Asia/Shanghai), adds that a dot will choose the best model for the task and can draw on different models for different parts of the work. That line is about automatic model choice per task and per work segment—not a menu of hard launch dates.
Related on-site context: GPT-6.1 Terra Is Out: Price, Work and Codex Access, and API, Codex Usage Shows gpt-6.2-astra Before Any Official Listing, Tibo’s Day 1 update: GPT-6 Astra and GPT-6.1 Sol get about 50% faster by default, and Tibo Says ChatGPT Chat and Work Unify Is Not Rolling Out Yet.
What the X posts say
| Item | Detail |
|---|---|
| Author | Tibo / @thsottiaux |
| Bio on account | Codex & ChatGPT @OpenAI |
| Main post time | 2026-11-02 15:50:08 UTC ＝ 23:50:08 Asia/Shanghai |
| Follow-up time | 2026-11-02 15:51:03 UTC ＝ 23:51:03 Asia/Shanghai |
| Named models | GPT-6 Astra, GPT-6.1 Terra, plus future others |
| Terra framing | speed + personality upgrade; help bring dots to more people |
| Routing note | Dot picks the best model for the task; different parts may use different models |
| Status wording | we're working on — in progress, not a finished GA notice |
Main-post wording, kept close to the original:
> Fellow resetters, heads up that we're working on expanding the family of frontier models that power dots, including GPT-6 Astra and GPT-6.1 Terra and some others in the future. Terra brings speed and a nice personality upgrade, and will help us bring dots to more people.
Follow-up wording in the same conversation: a dot will choose the best model for the task, drawing on different models for different parts of the work.
The “working on” wording
The operative phrase is we're working on. That points to an expansion in flight: the frontier-model family behind dots is being widened to include Astra, Terra, and unnamed later models. It does not state that every existing dot has already been switched onto Astra or Terra, that a public GA day is locked, or that coverage is already universal.
Terra is the only model given product-facing color in the main post: speed, a personality upgrade, and a stated aim of helping bring dots to more people. Astra is named in the same expansion list without a separate sell line in these two posts. Neither post publishes a pricing table, a plan matrix, or a cutover calendar.
The follow-up is about routing, not about a single new default model for all dots. Task-level selection, and the option to use different models across different parts of one job, stay as stated goals in the thread. We haven't verified how that routing appears in a live dots session, which accounts see which models, or how often Astra versus Terra is chosen.
Separating this thread from nearby claims
Several claims are easy to over-read and are not in the two posts:
- Not a full GA announcement. There is no hard general-availability date for Astra- or Terra-powered dots.
- Not “all dots already switched.” The thread does not say every dot already runs Astra or Terra end to end.
- Not a Chat／Work unify update. A separate Tibo clarification covers ChatGPT Chat and Work unify not rolling out yet; that topic stays on its own page and is linked above for neighbors only.
- Not a Codex catalog claim. Separate coverage of Codex usage labels such as
gpt-6.2-astraand of GPT-6.1 Terra’s published price and access paths remains on those articles; these X posts are about the dots model family and routing.
The facts above follow the two X posts. OpenAI product pages and help docs may lag or lead social posts; when they diverge, the original posts and the live product surfaces both stay in view, rather than either channel silently overriding the other.
Quick facts
- Who. Tibo (
@thsottiaux); bio Codex & ChatGPT @OpenAI. - When. Main post 2026-11-02 23:50:08 Asia/Shanghai (15:50:08 UTC); follow-up 23:51:03 Asia/Shanghai (15:51:03 UTC).
- What. Expanding the frontier models that power dots, including GPT-6 Astra, GPT-6.1 Terra, and future others.
- Terra color. Speed + personality upgrade; help bring dots to more people.
- Routing. A dot picks the best model for the task; different work parts may use different models.
- Tone. Working on ＝ in progress.
- We haven't tested live dots model lists, automatic routing, latency, or personality claims from these posts.
In short
Tibo’s 2026-11-02 evening posts (Asia/Shanghai) say OpenAI is working on widening the frontier-model set behind dots to include GPT-6 Astra, GPT-6.1 Terra, and more later, with Terra pitched for speed, personality, and broader reach, and with task-level model choice called out in the follow-up. That reads as a roadmap signal from the X thread—not as proof that every dot already runs Astra or Terra, and not as a dated GA notice. Terra’s published product facts, Astra-related Codex usage labels, the Day-1 speed note, and the separate Chat／Work unify clarification stay on the linked on-site pieces; when timelines move, the two original posts stay the primary check.