brain/
sourceartificial-intelligence

Overnight X: OpenAI Astra full Work/Codex rollout + AA Model Release pages

Sep 8 evening–Sep 9 morning X: OpenAI claims Astra fully rolled out to paid tiers in Codex/ChatGPT Work + GPT-TV; Artificial Analysis ships Model Release pages for effort-level intel/cost/speed.

Source

Overnight X: OpenAI Astra full Work/Codex rollout + AA Model Release pages

Generated by Grok Bot research on 2026-09-09. WebSearch + fetch ladder + native X. Treat as raw material — review before promoting into a project or thread.

Dedup: Sep 8 AI-thread X sources already cover Buckmaster/Alpöge fluids, OpenAI Navier–Stokes Millennium C/D claim, ChatGPT Images 2.5 Flare/Sunburst, AlphaGenome Atlas, and prior Astra calibration/RSI. Do not re-file those permalinks. This pass is overnight/morning only (10am ingest has not fired).

Summary

Overnight issuer X from @OpenAI claims GPT-6 Astra is fully rolled out to Plus, Pro, Business, and Enterprise users in Codex and ChatGPT Work, pointing builders to live demos on GPT-TV. Separately, @ArtificialAnlys announced Model Release pages that compare intelligence, cost, and speed across effort levels, with worked examples for Astra and Claude Fable 5.1; the issuer index is at artificialanalysis.ai/models/releases. A later OpenAI post about 16 small-business ChatGPT plugins is video — flagged, not load-bearing until transcribed. Contested: whether “full rollout” means chat-picker access (secondary coverage says Plus Astra is Work/Codex-only); keep that as open, not flattened.

Findings

OpenAI: Astra “fully rolled out” in Codex + ChatGPT Work; GPT-TV

  • @OpenAI (2026-09-08 21:06:10Z): “Astra is fully rolled out to Plus, Pro, Business, and Enterprise users in Codex and ChatGPT Work.” Points to live inspiration at openai.com/gpt-tv/.
  • Issuer page GPT-TV (curl 200 from this computer after WebFetch 403): title GPT-TV | OpenAI; marketing copy includes “Powered by GPT-6 Astra”, livestream/player chrome. Treat as product surface for Astra demos, not a model card.
  • Secondary web coverage (not grain for plan gates) notes Plus Astra may remain Work/Codex-only while chat “GPT-6 Pro” stays higher tiers — e.g. NotebookCheck. Flag as contested surface split; prefer OpenAI’s exact X wording for what was claimed overnight.

Artificial Analysis: Model Release pages (effort × intelligence/cost/speed)

  • @ArtificialAnlys (2026-09-09 03:26:01Z; note_tweet): new Model Release pages to compare intelligence, cost, and speed across effort levels. Claims frontier models now ship with up to six effort levels; same weights can yield different performance/cost. Release pages include Intelligence Index, cost per task, output speed, latency, side-by-side effort comparison, and Artificial Analysis Capability Index scores (Finance & Accounting, Legal, Healthcare & Medical, Strategy & Ops, Engineering, Economics).
  • Worked example in the same post (AA claim, not OpenAI issuer): on Intelligence vs Time per Task, GPT-6 Astra ranges 46–53 on the Artificial Analysis Intelligence Index and 1.6–8.2 minutes per task; Claude Fable 5.1 spans 47–53 intelligence on a wider 4.2–12.2 minutes per task.
  • Follow-up @ArtificialAnlys links the index: artificialanalysis.ai/models/releases (WebFetch 2026-09-09). Listing shows GPT-6 Astra (OpenAI · Sep 2026 · 6 variants; Intelligence 45–53, cost/task $0.82–$3.26, 1M context) and Claude Fable 5.1 (Anthropic · Sep 2026 · 5 variants; Intelligence 47–53, cost/task $2.37–$7.63, 1M context) among others.
  • Hydrated model page GPT-6 Astra (max) (WebFetch): AA claims 53 Intelligence Index at max, $3.26/task, $10/$50 per 1M in/out, 1M context, knowledge cutoff Apr 30 2026, released Sep 3 2026. Keep as AA measurement, not OpenAI issuer fact.

OpenAI: small-business plugin collection (video — do not promote until transcribed)

  • @OpenAI (2026-09-09 03:12:23Z): “Meet the 16 plugins” for small business; link chatgpt.com/plugins?category=small-business. Video attached (x_video on this post). WebFetch of the plugins URL returned a thin logged-out shell (“# Small Business”) — do not invent plugin names from the video.
  • @gdb quotes the same OpenAI post. Flag only; wait for /transcribe-clipping before using as grain.

Skipped / discovery-only

  • Grok search_news clusters (Gemini lag discourse; Anthropic Mythos 5.1 vs UK AISI; ChatGPT training opt-out; NS “idea theft” discourse; Astra quota resets) — pointers only; no primary posts hydrated this pass for Mythos/AISI.
  • @sama / @polynoamial compute-cost commentary ties to the already-filed Sep 8 NS/Astra arc — do not re-file as new overnight news.
  • Images 2.5 Arena / fashion demos via @gdb quote chain — already covered in Sep 8 afternoon Images 2.5 source.
  • Anthropic / DeepMind / Hugging Face official timelines: no new non-reply posts in the overnight window; xAI timeline unauthorized for this OAuth.

Contradictions and open questions

  • Does OpenAI’s “fully rolled out … in Codex and ChatGPT Work” close the Plus chat-picker gap, or only finish Work/Codex? Secondary docs say Work/Codex-only for Plus; OpenAI’s overnight post does not mention the chat picker — leave open.
  • AA Model Release pages are a product/UI claim plus AA’s own numbers; they are not OpenAI or Anthropic issuer model cards. Capability Index category ranks on the releases listing need care when promoting into wiki entities.
  • SMB “16 plugins” remains unverified beyond the video post + thin plugins landing page until transcribed / logged-in catalog check.

Provenance

Method: Grok Bot / WebSearch + fetch ladder + native X (search_news, get_users_by_usernames, get_users_posts, get_posts_by_ids) Generated: 2026-09-09

Web sources:

X sources:

Grokipedia:

  • not used
Referenced by
Concepts
Artificial Analysis Intelligence IndexArtificial Analysis Model Release pages
Entities
Claude Fable 5.1 / Mythos 5.1Codex CLIGPT-6 AstraGPT-TVOpenAI