Openai pulled gpt-6.1 astra after it failed internal safety tests, then shipped the value model the…
openai pulled gpt-6.1 astra after it failed internal safety tests, then shipped the value model the next day. gpt-6.1 sol is live in codex at $2 per million input tokens, about a fifth of astra's price, with near-astra results on complex tasks.
the alignment review didn't stop the release, it just repriced it.
Context
Verified source: the Wall Street Journal (Sep 28, 2026) reports OpenAI scrapped the release of GPT-6.1 Astra, due in ChatGPT and Codex in October, over safety concerns raised in internal testing. OpenAI introduced GPT-6.1 Sol on Sep 29, 2026, which it says nearly matches Astra on agentic coding, computer use and professional work at one-fifth of Astra's standard input and output prices.
OpenAI lists Sol at $2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens.
The sequence (Astra cancelled Sep 28, Sol on Sep 29), the $2 input price and the one-fifth figure match the WSJ and OpenAI's own post. The WSJ says the model was pulled over safety concerns raised in testing, and the exact wording 'failed internal safety tests' was not seen in the sources read, so unsupported here, not refuted. in the excerpts read. OpenAI's addendum shows Sol had 28 flags at severity 3 or above in a simulated Codex deployment. 'The review repriced it' is the author's interpretation.
Related work
- GPT-6.1 Sol System Card, OpenAI ↗Dated Sep 29, 2026.
- GPT-6.1 Sol in GitHub Copilot, GitHub Changelog ↗Same-day availability.
Watch next
- Read the Sol system card addendum for what the safety review measured.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 30 September 2026 at 20:51 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →