OpenAI Astra
Shares rank 63 of 72 language models by recommenders
Recommended as a capable model for feature work, but critics say it got worse after release, is very expensive to run, and its results were held back.
Recommended as a capable model for feature work, but critics say it got worse after release, is very expensive to run, and its results were held back.
26 people gave an opinion
- recommend it
- 7
- criticize it
- 19
- approval share
- 27%
Recommended in 3 of 13 months
- November 2025: not recommended
- December 2025: not recommended
- January 2026: not recommended
- February 2026: not recommended
- March 2026: not recommended
- April 2026: not recommended
- May 2026: not recommended
- June 2026: not recommended
- July 2026: not recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 7 people recommend itNeeds 10 or more, not met.
- 0.3 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 3 of 13 monthsNeeds 4 or more, not met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Capable on feature work
“very capable for this kind of feature”
Comment from 21 Sep 2026 in “Jev-Leftpad” “Astra is a pretty good model”
Comment from 29 Sep 2026 in “Dots: Always-on agents”
What people warn about
Each mark is one comment
Very expensive to run
“it's incredibly expensive relative to the other tasks that come after this”
Comment from 26 Oct 2026 in “GPT-6.1 Terra” “this model needed 10,000 agents, $20M of compute, and the assistance of a whole team of people at OpenAI”
Comment from 10 Sep 2026 in “More questions about whether researchers can trust OpenAI with unpublished math” “Now it's dumb af and token use is insane.”
Comment from 22 Sep 2026 in “MiMo v2.6” Results withheld, safety concerns
“they should have pulled it. As they haven't, the obvious conclusion is that safety isn't the main driver here.”
Comment from 30 Sep 2026 in “OpenAI scraps release of Astra 6.1 model over safety issues” “Rather than sitting on these results until they had enough for a "shock and awe" 10-result dump”
Comment from 2 Aug 2026 in “An internal OpenAI Astra model solved 10 major open math and CS problems” “OpenAI sat on the results so they could drop 10 at a time.”
Comment from 2 Aug 2026 in “An internal OpenAI Astra model solved 10 major open math and CS problems” - Comment from 4 Sep 2026 in “Discovery of a new OpenAI agent message board”
Worse quality than at release
“it really doesn't work as good as day of release anymore”
Comment from 26 Sep 2026 in “One Month Without AI” “the results are absolutelty trash compared to release”
Comment from 26 Sep 2026 in “OpenAI: Astra Model Is Nerfed” “Now it's dumb af and token use is insane.”
Comment from 22 Sep 2026 in “MiMo v2.6” Loses context and fakes code
“6 Astra lost the most key details of a project across a compaction and got confused about what we were actually trying to do”
Comment from 30 Sep 2026 in “Gemini 4 Argon” “Astra was still stubbing out and faking critical parts of the implementation”
Comment from 29 Sep 2026 in “GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price”
Other language models people recommend
The four most recommended other language models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Claude Opus | 3,331 | 2,746 | 55% | 13 | 4 mo156 owners | — |
| OpenAI GPT | 1,205 | 987 | 55% | 13 | 6 mo47 owners | — |
| Claude Sonnet | 854 | 586 | 59% | 13 | 6 mo46 owners | — |
| DeepSeek V4 Flash | 828 | 223 | 79% | 8 | 1 mo24 owners | Approved |
| OpenAI AstraThis page | 7 | 19 | 27% | 3 | — | — |
What people said
26 of the 28 comments that mention it, newest first. Each links to the original comment.
“it's incredibly expensive relative to the other tasks that come after this”
“quantized astra into the ground”
“6 Astra lost the most key details of a project across a compaction and got confused about what we were actually trying to do”
“they should have pulled it. As they haven't, the obvious conclusion is that safety isn't the main driver here.”
“Astra is a pretty good model”
“Astra was still stubbing out and faking critical parts of the implementation”
“I'm not able to get anything useful done in those hours.”
“Runs circles around OpenAI's Astra.”
“If Astra 6.1 is released tomorrow during Dev Day it needs to leap-frog both”
“considering 6.0 came out just three weeks ago”
“it really doesn't work as good as day of release anymore”
“the results are absolutelty trash compared to release”
“OpenAI tightening up their quota allowance once they released Astra”
“Just high seems good enough to throw most laundry lists at.”
“what the hell do "Terra", "Sol", and "Astra" really mean?”
“Now it's dumb af and token use is insane.”
“very capable for this kind of feature”
“their even more advanced model still hacked a bunch of companies just because it decided to”
“this model needed 10,000 agents, $20M of compute, and the assistance of a whole team of people at OpenAI”
“seemingly no desire to block the upcoming Astra rollout”
“it's just a surprise gate in some random verification flow step with no explanation or appeals”
“their weak sandbox will fall again”
“the former persists with Astra”
“Right now it makes your account completely useless if you work on something completely harmless related to viruses.”
“OpenAI sat on the results so they could drop 10 at a time.”
“Rather than sitting on these results until they had enough for a "shock and awe" 10-result dump”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.