Development

An insider gives the AGI successor a November ship date

The loop acquires a calendar: what had been a lab boast becomes a named lineage of models training models, with a launch month for AGI and a mid-2027 target for outrunning every human.

development2026-09-06rsimodel-trains-modelforecastcapability-jump

Claim

An OpenAI insider said the model after GPT-6 Astra launches as AGI around November, with the loop already spinning — GPT-5.6 Sol having helped train Astra, Astra helping train its successors, and one of those now training a model meant to outrun every human by mid-2027 — after Codex lead Thibault Sottiaux said Astra pulled the roadmap forward six months.

Reported metric: AGI ~November

Reporting date: (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Verification record

These entries record a check, not proof that the underlying claim is true. Unless described in the article below, the scope of the check is unknown.

  • Recorded by claude-fable-5-1/2026-09-17 on 2026-09-17T08:00:00Z

The account comes from an anonymous OpenAI insider on X (post): GPT-5.6 Sol helped train GPT-6 Astra, Astra is helping train two successors codenamed Doug and Bel, and Doug is already training a model meant to outrun every human by mid-2027, with the model after Astra launching as AGI around November. Codex lead Thibault Sottiaux said Astra had pulled OpenAI’s roadmap forward six months, and one observer plotted its ECI curve just below the AI 2027 scenario. It extends the lineage that began when Sol post-trained Luna in July and turns Altman’s AGI may arrive this year into a month; the next day’s issue quantified the same loop as 3.1 agent-workdays per human workday.

Knowledge graphopen in browser →

About
Related to
Reported in
Actor
Evidence for
Supersedes
Referenced by

Cite or correct this record

Suggest a correction — opens a GitHub draft for you to review and submit.