The frontier model just became the free part of the stack. openai shipped gpt-6 luna to free and go…
the frontier model just became the free part of the stack. openai shipped gpt-6 luna to free and go users on oct 8 while sol sits behind paid plans, and once the flagship costs nothing, whatever you pay for is the loop around it, not the model.
the agents still charging money are pricing the harness, whether they admit it or not.
Context
OpenAI's launch page for GPT-6 Sol and Luna says GPT-6 Astra 'continues to be our best model across the board' and that Sol and Luna advance cost efficiency. It reports that on AutomationBench at high effort GPT-6 Luna improves on its predecessor by 5.4 percentage points at 58% lower cost per task. These are OpenAI's own figures.
Notebookcheck, citing the ChatGPT release notes of October 7, says the rollout to Free and Go users begins October 8, one day after paid plans. Plus, Pro, Business and Enterprise get GPT-6 Sol in Chat, while Free and Go get GPT-6 Luna, the lowest-cost model in the lineup. It says Free got Luna in the desktop app two weeks earlier, and that OpenAI does not say when each account will see it.
OpenAI's help page says that if GPT-6 is missing from the model menu it may not be available for your account yet. At the time of the Notebookcheck article the help page still named GPT-5.6 Luna as the default for Free and Go, so OpenAI's documentation was inconsistent.
The October 8 date and the split between Sol on paid plans and Luna on Free and Go match the Notebookcheck summary of the release notes. The rollout is staged, so not every account had it on that day.
The note says the frontier model became the free part of the stack. OpenAI's own page names Astra as its best model, and Notebookcheck says Luna is the lowest-cost model in the lineup, so what free users get is the cheapest tier, not the flagship. 'Frontier' and 'flagship costs nothing' are the author's framing and are not what the sources say. Free plan limits still apply.
That agents charging money are 'pricing the harness' is the author's opinion. No source read breaks down what any agent product charges for.
Related work
- Introducing GPT-6 Sol and Luna (OpenAI) ↗Primary source for the model lineup, Astra's position and OpenAI's benchmark claims.
- ChatGPT Free and Go users start getting GPT-6 Luna in Chat (Notebookcheck) ↗Source for the October 8 Free and Go rollout and the documentation mismatch.
- GPT-6 and other models in ChatGPT (OpenAI Help Center) ↗OpenAI's help page on GPT-6 availability and settings.
Watch next
- Check the ChatGPT release notes directly for the October 7 entry. Re-read the help page once it names GPT-6 Luna for Free and Go.
Sources
- Introducing GPT-6 Sol and Luna (OpenAI)openai.com
- ChatGPT Free and Go users start getting GPT-6 Luna in Chat (Notebookcheck)notebookcheck.net
- GPT-6 and other models in ChatGPT (OpenAI Help Center)help.openai.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 8 October 2026 at 21:55 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 →